Eo

Science and technology education, 44
The influence of computers

and informatics on
mathematics and its teaching
Edited by
Professor Bernard Cornu
Professor Anthony Ralston
UNESCO
Education Sector Paris
ED-92JWSll7 October 1992
The designations employed and the presentation of
the material in this document do not imply the
expression of any opinion whatsoever on the part of
UNESCO concerning the legal status of any country,
territory, city or area or of its authorities, or
concerning the delimitation of its frontiers or boundaries.
Published in 1992 by the United Nations Educational,

Scientific and Cultural Organization,
7 Place de Fontenoy,
75352 Paris 07 SP
Printed by UNESCO
0 UNESCO 1992
Printed in France
Preface
The three-way interaction between mathematica. One might conclude that school uses of com-
ics, computers and mathematics education is be- puters are only known in those two regions. This
coming stronger each year. How schools and univer- is certainly not true. But more abundant finan-
sities should respond is still an open question. This cial resources have permitted a greater penetration
document has been prepared to contribute to the of computers in schools and universities in Europe
debate. The following quotation from the Overview and North America than elsewhere. For the situ-
chapter states succinctly why this debate is so im- ation in the rest of the world, see the reference,
portant: An International Perspective, by Jacobsen in the
We are facing a situation in which children Annotated References.
are taught to do mathematics in ways that Unesco wishes to express its appreciation to the
are very largely outmoded, with at least 80% editors, Professors Anthony Ralston and Bernard
of curriculum time wasted on trying, more or Cornu, to the authors for their contributions, to
less successfully, to develop fluency in skills Professor Ralston for preparing the final manuscript
of now limit,ed value. and to Cambridge University Press for giving its
permission for UNESCO to include in this docu-
The International Commission on Mathemati-
ment some updated contributions from the original
cal Instruction (ICMI) undertook a study, The In-
publication.
fluence of Computers and Informatics on Mathe-
The views expressed in this report are those of
matics and Its Teaching, which included a con-
the editors or the individual authors and not neces-
ference in Strasbourg, France in 1985, in which
sarily those of UNESCO.
UNESCO co-operated. The outcome of the Study
We welcome comments on the contents of this
was a book published by Cambridge University
document, which should be sent to: Mathemat-
Press, bearing the conference title. With the
ics Education Programme Specialist (Science and
quick pace of change of computers, mathemat-
Environmental Education Section) or Mathemat-
ics and its teaching, the books contents have be-
ics and Computing Programme Specialist (Basic
come outdated. The development of this new
Sciences Division), UNESCO, Place de Fontenoy,
document is explained in the Editors Foreword.
75700 Paris, France.
The reader will notice that the authors of this
document are all from Europe and North Amer-
UNESCO, Paris
Editors Foreword to Second Edition
In 1985 the International Commission on Math-

ematical Instruction (ICMI) chose The Influence of For this edition the report of the Strasbourg
Computers and Informatics on Mathematics and Its meeting itself has been brought up to date by the
Teaching as the topic of the first of a series of stud- leaders of the three workshops held at that meeting
ies on topics of current interest within mathematics and five of the articles in the first edition have been
education. ICMI could not have chosen a more apt updated for this edition. In addition, the editors
and important topic. In the seven years since the have solicited four new articles written just for this
publication of the first edition of this book, the im- edition. The result, we hope, is a volume which will
portance of calculators and computers has grown be as well received as was the first edition and which
rapidly and is now the single most important factor will be useful to mathematics educators throughout
in creating change in all aspects of mathematics ed- the world.
ucation. Thus, ICMI with the cooperation of UN- Of course, the nature of computer and calcula-
ESCO asked us to edit an updated version of the tor technology is that it changes so rapidly - as do
original book which would retain the strengths of its implications for mathematics education - that
that volume but would also bring the topics in it up a third edition will no doubt be needed in several
to date and, as well, incorporate topics which were years. But the current volume gives a fair picture
not adequately discussed in the first edition but are of the impact of computers and calculators on math-
now of major importance (e.g. symbolic mathemat- ematics education in 1992. It should, therefore, pro-
ical systems, algorithms and algorithmics). vide a valuable resource for mathematics educators
The conference at Strasbourg in 1985 whose pro- who wish to learn about this impact or who wish to
ceedings were incorporated in the first edition was incorporate the technology into their mathematics
organized by a Program Committee consisting of R. teaching or their teaching of prospective mathemat-
F. Churchhouse (Cardiff), B. Cornu (Grenoble), A. ics teachers.
P. Ershov (Novosibirsk), A. G. Howson (Southamp- The editors wish to thank UNESCO for its sup-
ton), J.-P. Kahane (Orsay), J. H. van Lint (Eind- port for this project, particularly Angelo Marzollo
hoven), F. Pluvinage (Strasbourg), A. Ralston (Buf- and Edward Jacobsen.
falo) and M. Yamaguti (Kyoto). The proceedings This book was produced using TeX at the State
were edited by A. G. Howson and J.-P. Kahane and University of New York at Buffalo.
were published by Cambridge University Press.
Bernard Cornu
Anthony Ralston
June 1992
...
111
_I_-- -_^,. p . ..- -..

CONTENTS
H. Burkhardt and R. Fraser: Overview 1
An Update of the 1985 Strasbourg conference: 11

R. F. Churchhouse: The Effect of Computers on Mathematics 12
A. Ralston: The Impact of Computers and Computer Science on
the Mathematics Curriculum 19
B. Cornu: Computers as an Aid to Teaching and Learning
Mathematics 25
L. A. Steen: Living with a New Mathematical Species 33
S. B. Maurer: What are Algorithms? What Is Algorithmics? 39
J. Stern: On the Mathematical Basis of Computer Science 51
K.-D. Graf, R. Fraser, L. Klingen, J. Stewart and B. Winkelmann:

The Effect of Computers on the School Mathematics Curriculum 57
S. B. Seidman and M. D. Rice: A Fundamental Course in Higher Mathematics

Incorporating Discrete and Continuous Themes 80
B. Cornu: Teacher Education and Training 87
B. R. Hodgson and E. R. Muller: The Impact of Symbolic Mathematical

Systems on Mathematics Education 93
M. Mascarello and B. Winkelmann: Calculus Teaching and the Computer.

On the Interplay of Discrete Numerical Methods and Calculus
in the Education of Users of Mathematics 108
D. Tall and B. West: Graphic Insight into Mathematical Concepts 117
Annotated References 124
Index 131
---,-._ 1.L ..*

.-*.._
--
AN OVERVIEW
Hugh Burkhardt
University of Nottingham, U.K.
Rosemary Fraser
The challenge
The overall picture
Where are we going? Where do we want to go?
Why? How do we know? How may we find out It may be useful to begin with an overview of
more? How do we get it to happen? the present situation in three separate domains of
As far as the influence of computers and infor- activity:
matics on mathematics and the mathematics cur- Doing Mathematics - this is the domain of math-
riculum is concerned, these are the central questions ematical activity; in every sphere, from everyday
that this volume, like its predecessor, will address. uses to research, it has been revolutionised by
We shall also be concerned with progress in the technology.
seven years since the original Strasbourg meeting. Understanding of the Learning and Teaching of
Which aspects have moved quickly and substantially Mathematics - this domain is concerned with
towards reasonably firm conclusions? On which ar- the processes of learning and teaching con-
eas is the situation now little different from then? cepts, skills and strategies in mathematics and
What needs to be done about it? its applications; it is clear that technology has
A mismatch of timescales is one of the central profound implications here, both through the
challenges of this field which does not normally oc- changes in doing mathematics and as a potential
cur in the processes of change, either within mathe- aid to learning and teaching, but these phenom-
matics itself or in the development of new curricula. ena are not yet well-understood.
The pace of change in the technology is much faster Mathematics Curricula and Teacher Training -
than has ever been achieved for school curricula; both the first two domains have implications for
typical timescales for significant changes to occur curricula, including both materials and teacher
are roughly as follows: support; the development of new curricula that
reflect the changed learning objectives and use
computer technology a few years technology effectively in their realisation is a ma-
mathematics research 10 - 20 years jor task.
school curricula 5 - 20 years The pattern of change so far is summarised in
Table 1 on the next page.
Thus we should be aware that when we design new
We shall now discuss each of the three domains
curricula to use the power of new technology, we
in more detail.
shall continually be behind the times. This moving
target problem is well recognised but needs to be
Changes in doing mathematics
addressed at a strategic level in planning change. If
the new curriculum elements are to be robust and In Domain A we now have a situation in which
widely useful, the curriculum designer cannot as- the changes in the way mathematics is done, at ev-
sume a specific level of technological provision and ery level from the shopkeeper to the research mathe-
sophistication in schools - both will vary widely matician and engineer, are moving purposefully for-
from time to time and from place to place. ward with the advances in the technology, and with
This is important. If each student has a micro, the methods for its utilisation that informatics helps
curriculum possibilities open up which are not there to develop. Obviously there is some time lag but,
with one micro per class; even these possibilities de- at least in comparison with the exploitation within
pend on the sophistication of the micro - one line mathematical education, there is no serious imped-
of display, a few lines, many lines, graphics, access iment to change. The reasons are fairly clear
to data - each step is significant. Equally, it is al- l those involved have a clear incentive to use the
ready clear that even quite low levels of computer new methods, which give them more power with
provision and sophistication still have enormous ed: less pain
ucational potential. Is technical restraint a virtue, l and at relatively modest cost that is more than
or does it impede progress? made up for in increased effectiveness.
2 Influence of Computers and Informatics on Mathematics and Its Teaching
are now numerical and, particularly, graphical, al-

A Doing B Learning C System lowing an experimental approach with much less ef-
Mathematics and Teaching Change fort than in the past. Within mathematics most
areas have been affected: discrete mathematics and
Where to combinatorics, number theory, algebraic and differ-
go? Why? ential equations, finite groups, fractals and chaos, as
well as all aspects of data analysis - these have all
lots of ideas lots of ideas + very little been profoundly changed; the papers in this volume
+ well tried a little exper- sign; even cal- describe some examples. The role of algorithmics
experience ience growing culators not is now central (see the chapter by Maurer). But
in patches integrated equally clear is the effect on other aspects of pure
mathematics - even the definitions of elegance and
How do the status of proof.
we know? Realistic applications to practical situations
have suddenly become more accessible as the
many examples many effect- observation drudgery associated with realistic numbers and
work well ive examples more realistic models is cut away. No longer is the
focus on extracting the maximum from the few ana-
How do we lytic models that are tractable by traditional meth-
find out ods.
more? Indeed, the role of such models in the future is
an important issue for both mathematics and the
it will happen systematic enrichment curriculum. It seems likely they will continue to be
because par- small-scale packages centrally important, not as methods of solving prob-
ticipants have studies in + lems, but as vivid illustrations of important effects.
incentives realistic cir- whole ex- A closed algebraic expression displays, for those who
cumstances perimental can read it, the dependence on all the variables in
+ curricula the model - something which, if there are many vari-
performance ables, can be very difficult to communicate graph-
measures ically or numerically. (For example, the expression
for the response of a damped harmonic oscillator to
How do we a sinusoidal driving force depends on five variables
get it to - understanding the phenomenon of resonance from
happen? numerical solutions alone is not easy).
Several of the chapters that follow are focussed
it will happen more support study dynam-
on the mathematical issues. Churchhouses review
because par- for system- ics of change
ranges over the various fields and aspects of mathe-
ticipants have atic r & d + matical activity, looking at the effects of technology
incentives model ex-
on the way mathematics is being done. Stern is
periments
mainly concerned with the impact of computer sci-
in realis-
ence rather than technology on the way mathemat-
tic circum-
ics is done while Steen introduces the perspective
stances
of the computer as a new mathematical species.
Ma,urer addresses what is, perhaps, the central area
Table 1 of change - the dominant place of algorithmic think-
This progress is set out in some detail in the ing and its implications.
present volume, not only in the introductory chap- Finally, a word of warning on the student as
ters, but in the contributions of Steen and Stern, mathematician. Largely because of the imitative
and to a considerable extent in the other articles. nature of the current curriculum, it is easy to get
We shall therefore review it briefly and in general a quite false picture of students capabilities. A ma-
terms - the other chapters bring these generalities ture mathematician has command of a range of con-
to life. cepts and techniques (or knows where and how to
The changes in mathematics pervade the sub- get such command) and uses them autonomously to
ject. The methods of first recourse in many areas express and manipulate ideas and relationships to
Overview 3
get answers and understanding. There is clear evi- Thus rigour and vigilance are needed in this devel-
dence that, on such criteria, students autonomous opment process.
performance is several years at least behind their In case there are any who believe that we exag-
performance on imitative exercises. The calculator gerate the dangers, let me draw attention to a few
is a useful resource because teenage students can famous examples of intended innovations in mathe-
already use arithmetic for a range of purposes; in matical education which turned into something en-
contrast it has been shown, for example, that even tirely different:
very bright 17 year-old students may not use algebra
The splendid Bourbaki enterprise was
at all as an autonomous mode of expression, though
launched (believe it or not (Weil, 1979)) to
they have had 5 years of success in manipulating
establish a firmer foundation for mathemat-
it (Treilibs et al, 1981); so, for example, the bene-
ical education; few now see that as among
fits of a machine that will manipulate in a language
the positive contributions it has made, while
they do not speak fluently are elusive, and maybe
many are concerned at the effects of overem-
illusory.
phasis on formalism that has arisen from this
The overall effect of these changes is well- approach in school mathematical education.
summarised by Mascarello and Winkelmann, in a
way that clarifies the challenge to designers of cur- Smalltalk was originally devised by the Xe-
riculum: rox Learning Research Group largely to pro-
duce a medium, the Dynabook, that would
In total, there can be observed a specific be as natural to a child as pencil and pa-
shift in the spectrum of abilities, from pre- per (Goldberg, 1978); what has emerged is
cise algorithmic abilities to more complex in- perhaps the most sophisticated graphics ori-
terpretations, so to speak from calculation entated data management system so far -
to meaning, which in a certain sense is a re- an important achievement, but a very dif-
versal of the historical evolution. In this pro-
ferent thing. (The Learning Research Group
cess the mathematics to be mastered tends to was renamed the Software Concepts Group.)
become intellectually more challenging, but Smalltalk has not, at any rate, done any
technically simpler. harm to the school curriculum, and among
its offspring, the Macintosh microcomputer,
Changes in mathematics education may yet contribute notably in a quite differ-
ent way.
As to the other Domains, B and C, nearly all the
chapters that follow make suggestions for new cur- Our final example must be the reform move-
riculum elements based on these new methods of do- ment in mathematical education of 39 years
ing mathematics; readers will find many of these ar- ago - new math , modern mathematics and
guments stimulating, and even persuasive. Changes so on. Comparison of the initial aims agreed
are surely needed and these suggestions seem better at conferences, the pilot schemes in a few ex-
grounded than most. ceptional schools, and the classroom reality
Nonetheless, it must be recognised that such sug- of today shows the contrasts vividly. For ex-
gestions are fundamentally speculative at the level of ample, in England the applications of math-
ematics occupied a central place in the origi-
large-scale implementation - by which we mean that
nal design; in most of the major courses that
converting them into a well-developed and tested
curriculum for the typical teacher and the typical emerged applications were mentioned only
student is still a major challenge. This is the task to illustrate techniques with no serious at-
of Domains B and C. We can have no reliable idea tention to the practical situations involved.
Equally, new mathematical concepts were in-
how far any suggestions we put forward will prove
troduced but often with none of the payoff
feasible in any, let alone every, educational system.
that motivated their inclusion - because the
Even if they are implemented reasonably faithfully,
the full curriculum reality of what occurs will con- serious examples originally envisaged proved
tain many surprising side effects; more likely, the too difficult for most students, and were re-
placed with trivial ones. The second wave of
translation from an idea to a small scale pilot exper-
reform over the last decade has been rather
iment with exceptional teachers and facilities, and
more successful in remedying some of these
then to large scale reality will involve critical disi
tortions of the aims of the exercise which may even, defects, but has left others untouched.
in the end, call into quest.ion its curriculum value. These are cautionary exa.mples to bear in mind
._ -- v-e- .+Aeaw ..--.-- ._ .-- -. - .__--_ _~-_-... ...-~-

4 Inff uence of Computers and Informatics on Mathematics and Its Teaching
when looking at new possibilities; it is not easy to re- outside the classroom. Fig. 1, for example, shows
alise the potential vitality of a field of mathematics pre-, post- and delayed-test results of individual stu-
in large-scale curriculum implementation. (One can dents from comparable groups taught by two differ-
easily imagine a vivid commercial for the impor- ent methods. The first positive only method (a) is
tance and excitement of a new field like calculus; traditional explanation and reinforcement by prac-
compare it with the reality of a typical introductory tice; the second conflict method (b) is based on
American college text.) That does not mean that students discussion and debugging of errors gen-
vivid, effective implementation cannot be achieved erated by them from their own misconceptions. The
- but success requires high-quality engineering - greatly improved long-term learning is stable across
and some poetry as well as the facts. (We have ar- different topics (Bell and Basford 1989).
gued that, if the English language curriculum were score
like most mathematics curricula, the readings would 3
II
be drawn entirely from the telephone directory.)
A.
.
On the learning and teaching of mathematics
.
n
Progress in Domain B continues at a steady pace .
but, we would suggest, far too slowly to provide a
sound comprehensive underpinning for the new cur-
4
ricula that we need now. Some of this progress is
described in later chapters in this volume. These
developments and the associated research represent .
1
deeper insights into the way technology can affect
,
and enhance learning and teaching, together with 1
some elements of curriculum that can be and have
been used successfully in classrooms. They have ,
.
rarely been tested on a large scale and thus repre- I
sent only firm steps along the road towards the new
II
curricula (Domain C). Let us begin by looking at I
POST DEL
some general effects in technology-related change. PRE
(2)
In looking at curriculum reform, the first thing
to note is the scale of it - perhaps 80% of current
school classroom time is devoted to seeking fluency
in a range of pencil-and-paper technical skills, all of
which are now best done on computers of one kind
or another. This we call The Big Hole.
Secondly, the swing towards teaching mathemat-
ics that is intellectually more challenging, but tech-
nically simpler takes both teachers and curriculum
designers into areas outside the basis of their ex-
perience - thus such curriculum design should be
essentially a research-baaed exercise, if it is to work
well. It relates not only to content but to learning
and teaching style. Everywhere the curriculum is
still based on student imitation (e.g. HMI, 1977),
dominated by:
teacher explanation +
illustrative examples +
imitative exercises. IL
DEL
This can lead to rapid apparent student progress, PRE POST
but much research evidence (see Bell et al, 1983) (b)
shows that the skills acquired are not reliably re-

Figure 1
tained by most students, nor are they transferable
- particularly to non-routine problems in the world
Overview 5
To achieve the flexible competence of under- Equally, the challenge to explore must be at a
standing that the world requires, the pattern of level matched to the student - if the aim is to dis-
classroom activitieshas to be widened to include cover in an hour or so some important mathemat-
some which give more autonomy, more initiative to ical achievement that took a genius half-a-lifetime
the students. It is encouraging that the microcom- to create, the exploration will have to be so closely
puter has shown great promise in supporting such guided as to be essentially fake; on the other hand,
activities. interesting, though less global, problems which stu-
Thirdly, change is often threatening. Technology dents can tackle autonomously on their own re-
appears to reduce this threat, partly because it pro- sources, do exist at every level. For example, pro-
duces an obviously new situation and thus cannot gramming projects, at school and university, have
imply criticism of the teachers existing modes of op- shown some of the possibilities, and the difficulties
eration. This more than compensates for the extra for the teacher. A creative and systematic program
barrier of learning to use the equipment - provided of detailed empirical development will be essential if
it is reliable. Further, recent work on the teaching exploration is not to degenerate in most classrooms
of non-routine problem solving of a wide variety of into that closely guided discovery learning, which
kinds shows the importance of the strategic skills of is really an alternative style of explanation. The
comprehension, modelling, interpretation and eval- computer can, of course, help.
uation -just the skills that are brought to the front
of our attention by the computer.
I \ ILLUSTRATIONS OF /
In summary, the two great springs for change in APPLICATIONS
mathematical education in the past and the next
decade are technology and autonomy. Fortunately,
they can help each other, though there is much to tlATHEllATICAL
be learnt as to how best it might be done.
TOPIC
However, conversely, we believe that it may be
.-
important not to discount too easily the value of tra-
ditional skills, remembering that the current genera-
tion of innovators have the traditionalbackground,
as well as newly acquired skills with computers. Al- .-
most certainly, much of what was learnt is useless
but we need to check for losses as well as gains in HATHEHATICAL
AND OTHER SKILLS
a curriculum change. Mental facility with numbers,
graphs and expressions has always been an asset.
What is its status now?
Exploratory investigation as a key element in
the curriculum has been a major objective in En-
glish mathematical education for at least 30 years
- the Association of Teachers of Mathematics was
founded largely to promote it; in the USA, we know
that it has been a focus since Polya (1945) and offi-
I
cially central at least for a decade, since the NCTM
Figure 2
Yearbook on problem solving. However, despite
strenuous efforts it has not become a regular part The emphasis on problem solving encourages ap-
of the curriculum anywhere except in a tiny minor- plications of mathematics - even some with real
ity (less that 1 percent) of classrooms. We have a lot data. It is important to note two different kinds
of evidence and some understanding of how difficult of application, the illustrative and the situational
such activities are for the typical teacher to handle (see Fig. 2, from Swan 1990). In illustrative appli-
in the classroom; appropriate support must be de- cations the focus is really on the particular mathe-
veloped. Everyone rightly emphasises the curricu- matical topic; the applications are there in support -
lum opportunities for exploration, for experimental to help conceptual understanding through concrete
mathematics, that the computer provides; however, illustration, to show how mathematics can be ap-
the development of such an investigative element in plied, and to provide practice. In realistic, practi-
the curriculum will succeed only if it confronts the cal situations from outside mathematics the posi-
difficulties such activities present for teachers. tion is quite different - in principle, any or all of the
mathematics you know could help you to tackle the velopment .

problem, along with other knowledge and strategic
skills. Both these kinds of application are needed New mathematics curricula
- but the student must know which game is being
The lack of progress in Domain C is the major
played, because the best tactics are quite different. mismatch between intentions and outcomes over the
For illustrative applications the aim is to show how last seven years. It is notable that even the use of
much mathematics you know; for situations, it is simple calculators has not been fully integrated into
to provide the most powerful understanding of the the curriculum in any country in a way that realises
practical problem. their known potential for enhancing mathematical
Finally, assessment - in arriving at a curriculum, performance (even on traditional skills!).
assessment can be very helpful in clarifying cur- The reasons are less clear than is sometimes
riculum definitions, particularly by example. Whal thought by those who ascribe it simply to teacher in-
range of types of tasks do we want our students to ertia and/or parental opposition. Certainly, teach-
be able to do? It can be argued persuasively that a ers and other educators do not have the direct in-
task-defined curriculum has great advantages over a centives that the use of the technology provides
scope and sequence approach, though the two are for those doing mathematics. It does not so obvi-
complementary - one being synthetic and the other ously promise to increase their power as profession-
analytic. als, or to make their lives easier or more reward-
The chapters that follow range widely over this ing. Rather, it makes obsolete a large part of the
field, complementing analysis with the essential standard professional work of mathematics teachers
vivid exemplification which we have had to exclude and threatens them with, at least, a need for new
from this overview. Ralstons review takes a look skills, both mathematical and pedagogical. Equally
without prior assumptions (zero-based) at what parents and others tend to believe that their own
the curriculum should contain. The criteria include education remains valid - after all, look what it has
value to the student and, related, the way mathe- done for them!
matics is done nowadays together with lessons from However, when one compares the support offered
the psychology of learning. Cornus review covers a to teachers to make these changes with that which
wide range of roles for the computer in enhancing they routinely receive simply to sustain the current
teaching and learning - Computer-assisted Learn- curriculum, the contrast is stark. The textbooks
ing is as diverse as Paper-assisted Learning. The and other materials are not comparable, or often
remaining chapters that follow present a kaleid* even available. Retraining is sparse, as is coher-
scope of key aspects of computer use in learning and ent explanation of what is being attempted, and
teaching. Tall and West explore and illustrate the why. The temptation to blame the lack of change on
visual aspects of learning and the contributions that teachers is not only misguided but fruitless - they t
computer graphics can make. Hodgson and Muller are who they are. It is up to those who seek change
look at the other major shift enabled by the tech- to find, and to deliver, an effective and appropriate
nology - automated symbolic manipulation. Graf, mixture of pressure and support.
Fraser, Klingen, Stewart and Winkelmann take a We have evidence that teachers actually welcome
broader look at the various modes of use of tech- change, provided they are confident that the pace
nology in school and the potential of each in the and level of support is such that they can cope with
elementary school; both computers and calculators it without undue effort. As with many profound
are discussed. Steen takes a similar approach to curriculum changes, systems have so far failed to
college mathematics - where the issues range from provide any basis for such confidence.
the place of discrete mathematics to computer lit- In their chapter Cornu and Balacheff look at
eracy for all students. Seidman and Rice, and Mas- the problems of the new pedagogy and how it may
care110 and Winkelmann look in a most stimulating be communicated to teachers in training, a key
and practical way at a related central problem - the area in Domain C. We already have evidence (see
integration of discrete and continuous mathematics Burkhardt, 1984, 1985) that the potential of the mi-
within a college course. crocomputer for helping teachers to enhance student
All of these show what has been achieved in Do- learning presents a tremendous opportunity for cur-
main B - in realising, on a pilot scale, something riculum enhancement. The effects on the dynamics
of the enormous potential of computers and infor- of the classroom can be profound, but they are of-
matics in enhancing the learning and teaching of ten subtle; for this reason there is a great deal still
mathematics. They also set targets for future de- to do before we have even a broad understanding of
Overview 7
what can happen in the various modes of computer exploit the potential of computer support for the el-
use in education. ementary school mathematics classroom in a variety
We shall illustrate the sort of thing that may of powerful ways. Similarly, The Power Series (UC-
be expected by describing one application that has SMP/Shell Centre, 1992) offers an effective element
been developed and studied in some detail, and in teacher development, through the support that
which has proved particularly rich - the use by the single micro classroom can provide in explor-
the teacher of a single micro in the classroom, pro- ing new, more open ways of working. There are now
grammed to be a teaching assistant. We do so for many other enrichment materials using the com-
various reasons: It is less familiar to most people; puter to support learning, particularly those less-
it brings out some general points about the over- routine activities that many teachers find difficult
whelming importance of the people, teacher and to handle.
pupils, and of the dynamics of their interaction; Full technology-integrated curricula, with mate-
and it is particularly relevant to schools as we know rials to support them, are hardly available yet. If
them because it seeks to enhance the performance there is an exception, it is a few new courses in
of a teacher working with a group of children in the higher education, such as that described by Hodg-
classroom in the normal way. It also only requires son and Muller. There are early signs of moves
one microcomputer per class rather than one per to develop materials to support complete curricula;
child. some of the latest round of NSF-supported projects,
This mode of use, set out by one of us (Fraser, for example Seeing and Thinking Mathematically at
1981), has been shown to have remarkable effects in EDC and a parallel project at TERC, have a strong
leading typical teachers in a quite unforced and nat- emphasis on technology.
ural way to broaden their teaching style to include It is interesting that all three examples quoted
the open elements that are essential for teaching above (and many others) use the computer as a cat-
problem solving (Fraser et al, 1983, 1988). Since alyst for learning (Fraser, 1989) rather than as a
this is a crucial aim that reformers have been trying tool for doing mathematics or a tutorial system.
to achieve for at least thirty years with little or no No one doubts that the computer as a tool is a
effect, this is a valuable result. It is worth explaining central element in the curriculum but the develop-
briefly why these effects come about. First, the mi- ment of curricula to realise this is slow. Of course,
cro is viewed by the students as an independent per- the level of computer provision needed to make it
sonality. It takes over for a time a substantial part more than a passing experience is still beyond most
of the teachers normal load of explaining, manag- schools. It is interesting and ironic to remember the
ing, and task setting. These are key roles played by pioneering work of the Computer Assisted Mathe-
every mathematics teacher. The micro takes them matics Program (Johnson et al, 1966-68), in which
over in such a way that the teacher is led into less students learned mathematics in a Basic program-
directive roles, including crucial discussion with the ming environment; Kieren (1974) showed that far
children on how they are tackling the problem, pro- more students got through to fluency in algebra in
viding guidance only of a general strategic kind - this way - a result that has been confirmed but not
counselling if you like. yet implemented anywhere. As we have noted, even
These principles have been incorporated into the simple calculator is far from fully integrated into
a range of teaching materials that enable typical curricula.
teachers to sustain in their classrooms, without ex- Computer-based tutorial systems have contin-
ceptional effort, these learning activities of a more ued to emerge, and to become more sophisticated,
open kind. This book isjfull of other examples of sometimes embodying elements of artificial intelli-
curriculum components that have been shown to gence of an expert system kind. Apart from pro-
work well in typical circumstances, or can be devel- gramming itself, perhaps the first big idea for using
oped to do so. It is also important to recognise that computers in mathematical education was in teach-
there will be disappointments - or at least frustra- ing technical skills, particularly arithmetic. The ap-
tions - in the development process. In the last few proach followed the behaviourist teaching-machine
years there has been further progress to report in model. To provide effective teaching in this way has
the thorough, and imaginative, development of sub- proved a much harder problem than was expected.
stantial curriculum elements. They illustrate what We believe that it is still far from solution. It seems
can be done. The Journeys in Mathematics project that the computer-tutor can be effective in teaching
(EDC, 1991), funded by the National Science Foun- facts and straightforward techniques to people who
dation, has developed a series of modular units that have little difficulty with them; so, of course, are
other methods. However, despite great efforts by change?. It seems (Fullan, 1980) that both pres-
some talented people, it has not so far proved possi- sure and support are needed for effective change but
ble to write programs which are successful in diag- producing a balanced well-engineered package that
nosing and remediating students conceptual errors works is still an unsolved problem. One lever for
underlying technical skills that they find difficult. It change that cannot be ignored is the assessment sys-
remains true that tutorial systems have not begun tem; if a system does not recognise, measure and re-
to tackle the main defects of the traditional (and ward new curriculum elements then they will not be
largely current) mathematics curriculum, still con- taken seriously by many - WYTIWYG (what you
centrating on automating those learning activities test is what you get). Pressure is often preferred by
(largely drill-based) that are both over-represented politicians because it is less expensive than support,
and, on their own, ineffective. so that an effective balance is destroyed.
The questions we have raised imply a great deal
The next steps of work, integrating research techniques with cur-
riculum development, before we have even a basic
It seems from the above that Domains A and
understanding of the classroom potential that we
B are making steady progress and that Domain C
see so vividly illustrated in this book. Experience
presents the greatest difficulties - thus it seems that
suggests that, along the way, we shall find other
further work on large-scale implementation should
possibilities of at least as much promise.
become a priority over the next decade. There are
some small signs of movement in this direction in In order to realise any of these possibilities, they
various countries. However, the difficulty of achiev- will need to be systematically developed in detail
ing large scale change of any kind is often under- with representative samples of teachers and stu-
rated, or at least neglected. It clearly needs empiri- dents, using structured detailed data from the class-
cal study of the dynamics of change in the education room.
system as a whole, with all the factors this brings
in. We already know far more about the benefits Systematic research and development
that could flow from the use of technology (even
The slow progress in B and, particularly, C
within current financial constraints) than is realised
partly reflect a general problem in education - that
in practice. Without attention to Domain C, this
the level of expenditure on designing and developing
mismatch will simply get worse.
soundly-based changes is remarkably low; in Eng-
What, specifically, are we to do about this?
land and the US, for example, this research and
This is not the place for a serious discussion of
development ratio is substantially less than 0.1% of
methodologies of research and curriculum develop-
educational expenditure, whereas in other changing
ment (Burkhardt, Fraser and Ridgway, 1990). Very
fields such as medicine or modern industry it is typi-
briefly, there is no proven successful answer but
cally between 5 and 15%. We believe that this arises
some seem to be less susceptible to corruption of
because education is still dominated by the craft-
outcomes than others. We believe that the essence
based approach, which assumes that experienced
is an empirical approach - find out what actually
professionals have satisfactory methods of handling
happens to your draft ideas in practice, in circum-
each situation that presents itself - that everything
stances sufficiently representative of what you are
is basically well under control.
aiming for, and then revise the materials repeat-
This craft-based approach works well when two
edly until they work in the way intended. We have
conditions are satisfied:
found (e.g. Shell Centre, 1984) that such an ap-
proach, taken for granted in other fields, can com- the system is working satisfactorily
bine educational ambition with user-friendliness to a and
level not achievable with more casual development. there is no expectation of major change
Structured classroom observation makes a key con-
tribution to this approach, providing much richer Otherwise it involves the extrapolation of reliable
feedback than is often acquired in the development experience beyond its domain of validity - always a
of educational materials (Burkhardt, Fraser et al, hazardous process.
1982). However, more rigorous comparative evalu- In this respect the situation in education is
ation of alternative approaches is sorely needed. A rather similar to that in medicine a century ago,
few more comments are made below. or engineering further in the past. There are signs
It is important to ask of everyone in the sys- that the more systematic research-based approach
tem, but particularly teachers, Why should they (which now dominates the other two fields) is more
Overview 9
widely recognised as important in education, but mum)

there is still a long way to go. No one would dream C Curriculum change on a large scale - studies of
of using a drug, or flying in an aeroplane, that had how curriculum change can be effected and what
only been developed and tested as sketchily as most other school or social factors affect it? (lo4 - lo7
new (or, indeed, old) curricula. Nonetheless, the sit,- children minimum)
uation persists - perhaps because educational disas- All these levels are important, the earlier ones
ters are much less immediately visible. contribute to a fundamental understanding of the
Some dismiss such arguments on the grounds later ones. However, much more serious work has
that teaching and learning are much more varied been done at the early L and Tl levels; a more
and less controllable than engineering or medical balanced effort would be productive. The crucial
situations. Though this is true in some respects, distinction between Tl and T2 is often not made.
systematic observation of typical mathematics class- At the Tl level the teacher variables are almost ir-
rooms (HMI, 1979) shows a remarkable uniformity relevant - work there simply shows that, there is a
in delivering a curriculum that is inappropriate and teacher, usually a member of the development team,
seriously impoverished by current standards. Simi- who can make these things happen. At the T2 level,
lar results are observed in most countries. curriculum developers face the challenge of show-
The research and development ratio illustrates ing that a wide range of unexceptional teachers, in
how little serious attempt has been made in educa- normal circumstances of support, can also function
tion to work systematically to do better. The pro- in the desired ways. Similarly, the C level brings
cesses of systematic development are either absent in all the variables that relate to the pressures on
or sketchy, particularly in the quality of feedback the classroom from school and society, which are so
and the typicality of the development environment. critical to the implementation of any change. These
The search for better development methods is al- distinctions appear to be important in the limited
most non-existent. We believe that we are still es- impact of technology so far - and in doing better in
sentially leaping off cliffs, flapping wings tied to the future.
our shoulders - but nobody notices the mess.
These things need not be. So far from costing Acknowledgements
money, they represent a potential source of large
saving in the overall operation of the education sys- We have benefited from many conversations with
tem. out friends in the ITMA Collaboration, particularly
These considerations are not, of course, confined Richard Phillips, Jan Stewart, Jim Ridgway and Jon
to technology-inspired change. They are, however, Coupland, and others around the world, particularly
particularly acute in that circumstance because of David Tall and Tony Ralston.
the pace of change needed. We are facing a situation
in which children are taught to do mathematics in REFERENCES
ways that are very largely outmoded, with at least
80% of curriculum time wasted on trying, more or Bell A.W., Costello J. and Kuchemann D. [1983]:
less successfully, to develop fluency in skills of now- A Review of Research on Mathematical Educa-
limited value. tion, Part A: Research on Learning and Teach-
In planning a systematic approach, we think it is ing, Windsor: NFER- Nelson.
useful to distinguish different levels of research and
Bell A.W. and Basford D. [1989]: A conflict and in-
development in education. Studies at each succes- vestigation teaching method and an individu-
sive level involve an order of magnitude more stu- alised learning scheme - a comparative exper-
dents, and teachers, than at the previous one. Four iment on the teaching of fractions, Notting-
levels are: ham: Shell Centre for Mathematical Educa-
L Learning - studies of students learning, the tion; also Teaching for the Test, Times Educa-
nature of cognitive processes, difficulties and tional Supplement, 27 Oct.
misconceptions (10 - 10 children minimum)
Tl Teaching Possibilities - studies of different Burkhardt H., Fraser R. et al [1982]: Design and De-
kinds of stimuli and their effects on student velopment of Programs as Teaching Material,
learning (10 - lo2 children minimum) London: Council for Educational Technology.
T2 Realizable teaching - studies on what can a& Burkhardt H. [1984]: How can micros help
tually be achieved with typical teachers under in schools?: research evidence, Nottingham:
realistic circumstances (lo2 - lo3 children mini- Shell Centre for Mathemat,ical Education.
-* ___- .---- - .--. -_^- .

Burkhardt H. [1985]: The Microcomputer: Miracle Shell Centre [1984]: Problems with Patterns and
or Menace in Mathematical Education in Pro- Numbers, a Module of the Testing Strategic
ceedings of ICME5 (M. Carss, Ed.) Berlin: Skills Programme, Nottingham: Shell Centre
Birkhauser. for Mathematical Education.
Burkhardt H., Fraser R. and Ridgway J. [1990]: Swan M. [1990]: Mathematical Modelling for All
The Dynamics of Curriculum Change in De- Abilities in Proceedings of ICTMA4, Chi-
velopments in School Mathematics Around the chester, U.K.: Ellis Horwood.
World, Vol2 (I. Wirszup and R Streit, Eds.), Treilibs V., Burkhardt H. and Low B. [1981]:
Reston, VA: National Council of Teachers of Formulation Processes in Mathematical Mod-
Mathematics. elling, Nottingham: Shell Centre for Mathe-
EDC [1991]: J oumeys in Mathematics, Scotts Val- matical Education.
ley, CA: Wings for Learning. UCSMP/Shell Centre [1992]: The Power Series (six
Fraser R. [1981]: H ow to use Homo Sapiens in a modules for the single micro classroom in the
Computer Environment in Proceedings of the elementary school), Chicago: University of
1980 A DV Conference, Vienna: ADV; also Chicago School Mathematics Project and Not-
Design and evaluation of educational software tingham: Shell Centre for Mathematical Edu-
for group presentation in Microcomputers in cation.
Secondary Education (J. Howe and P. Ross, Weil, A. [1979]: History of Mathematics in Pro-
Eds.), London: Kogan Page. ceedings of the 1978 International Congress of
Fraser R., Burkhardt H., Coupland J., Phillips R., Mathematicians, Helsinki: ICM.
Pimm D. and Ridgway J. [1983, 19881: Learn-
ing Activities and Classroom Roles, Notting-
ham: Shell Centre for Mathematical Educa-
tion and J. Math. Behaviour, 6, 305-338.
Fraser R. [1989]: Introduction to Computers and the
Teaching of Mathematics (E. Dubinsky and R.
Fraser, Eds.), Nottingham: Shell Centre for
Mathematical Education.
Fullan M. [1980]: The Meaning of Educational
Change, New York: Teachers College Press.
Goldberg A. [1978]: Trends in Hardware and Soft-
ware in Informatics and Mathematics in Sec-
ondary Schools: Impacts and Relationships,
1977 IFIP Conference (D.C. Johnson and J.D.
Tinsley, Eds.), Amsterdam: North Holland.
HMI 119791: Aspects of Secondary Education in
England, Report of the HMI Secondary Sur-
vey, London: HMSO.
ICMI [1984]: The Influence of Computers and In-
formatics on Mathematics and its Teaching,
LEnsezgnement Mathmatique, 30, 159-172.
Johnson D.C. with Hatfield, L., Walther, J., La
Frenz, D., Katzman, P. and Kieren, T. [1966-
68]: Computer Assisted Mathematics Pro-
gram, Glenfield, IL: Scott Foresman.
Kieren T [1978]: Znformatics and the Secondary
School Mathematics Curriculum in Informat-
its and Mathematics in Secondary Schools:
Impacts and Relationships, 1977 IFIP Confer-
ence (D.C. Johnson and J.D. Tinsley, Eds.),
Amsterdam: North Holland.
An Update of the 1985 Strasbourg Conference
The first edition of this book grew out of a conference in Strasbourg in March 1985. The attendees at
that conference divided themselves into three Working Groups on the subjects of The Effect of Comput-
ers on Mathematics, The Impact of Computers and Computer Science on the Mathematics
Curriculum and Computers As an Aid to Teaching and Learning Mathematics. The reports of
these three working groups formed the first three chapters in the previous edition. In this edition the leaders
of the three workshops have updated the reports which appeared in the previous edition. These updated
reports appear on the following pages.
11
_,-.-- .~
-- ^ .-.-. l_-.-..- >uw,l
Part I
THE EFFECT OF COMPUTERS ON MATHEMATICS
R. F. Churchhouse
University of Wales, Cardiff, UK
1.0 Introduction
Mathematical concepts have always depended equations is the effective calculation and the qual-
on methods of calculation and methods of writing. itative study of solutions. The functions in which
Decimal numeration, the writing of symbols, the one is interested therefore are calculable functions
construction of tables of numerical values all pre- and no longer only those which are tabulated. The
ceded modern ideas of real number and of function. theories of approximation and of the superposition
Mathematicians calculated integrals, and made use of functions - developed well before computers - are
of the integration sign, long before the emergence of now validated. The field of elementary functions
Riemanns or Lebesgues concepts of the integral. In is extended, through the discretisation of nonlinear
a similar manner, one can expect the new methods problems. Informatics, too, compels us to take a
of calculation and of writing which computers and new look at the notion of a variable, and at the link
informatics offer to permit the emergence of new between symbol and value. This link is strongly
mathematical concepts. But, already today, they exploited in mathematics (for example, in the sym-
are pointing to the value of ideas and methods, old bolism of the calculus). In informatics, the necessity
or new, which do not command a place in contem- of working out, of realizing the values has presented
porary traditional mathematics. And they permit this problem in a new way. The symbolism of func-
and invite us to take a new look at the most traditions is not entirely transferable, and the attributes
tional ideas. of a variable are different in languages such as For-
Let us consider different ideas of a real number. tran, Lisp and Prolog.
There is a point on the line R, and this representa- In the sections that follow we look at some as-
tion can be effective for prompting the understand- pects of how computers and informatics have al-
ing of addition and multiplication. There is also an ready affected mathematics and mathematical re-
accumulation point of fractions, for example, con- search and present some thoughts on what future
tinued fractions giving the best approximation of a effects might be seen. We do not claim that our
real by rationals. There is also a nonterminating survey is comprehensive, especially so in the disci-
decimal expansion. There is also a number written plines of applicable mathematics, but we hope that
in floating-point notation. Experience with even a it provides some pointers. In any event information
simple pocket calculator can help validate the last technology, in the widest sense, is advancing far too
three aspects. The algorithm of continued fractions fast for any predictions to be of value for a period
- which is only that of Euclid - is again becoming a of more than a few years.
standard tool in many parts of mathematics. Com-
plicated operations (exponentiation, summation of 1.1 New and revived areas of mathematical
series, iterations) will, with the computers aid, be- research
come easy. Yet even these simplified operations will Computers not only provide a new tool in math-
give rise to new mathematical problems: for exam- ematical research and teaching. They are, at the
ple, summing terms in two different orders (starting same time, themselves the source of new areas of
with the largest or starting from the smallest) will research. Not all of the research stimulated by the
not always produce the same numerical result (see, availability of computers is in new branches of math-
e.g., Churchhouse, 1980, 1985). ematics; some is of ancient lineage, going back to the
Again, consider the notion of function. Teach- 19th or 18th century, but open now to attack with
ing distinguishes between, on the one hand, elemen- a weapon not available to Euler, Gauss, Jacobi, Ra-
tary and special functions - that is, those functions manujan, etc. Who can doubt, though, that these
tabulated from the 17th to the 19th century - and, giants of the past would have exploited these new
on the other, the general concept of function intro- possibilities with enthusiasm had they been avail-
duced by Dirichlet in 1830. Even today, to solve a able? It is one of the unique features of mathe-
differential equation is taken to mean reducing the matics that it is based upon a body of results that
solution to integrals, and if possible to elementary never loses its value. Fashions and interests may
functions. However, what is involved in functional change, but the neglected subject of the last cen-
12
Effect of Computers on Mathematics 13
tury, or even of the last millennium, may prove to and Piper, 1982). Many of these applications have
be of new interest at any time when conditions are benefited considerably from the availability of pro-
right for its re-emergence. So the corpus expands; gram packages specifically designed as an aid for re-
nothing ever dies, though it may remain dormant searchers in the field; the CAYLEY system for the
for centuries. In the age of information technology study of finite simple groups is a well-known exam-
we need to emphasize this fact, for it underlies ev- ple. Another is the development of Symbolic Math-
erything that follows. ematical Systems (see Section 1.6 and the chapter
One of the most famous examples of mathe- by Hodgson and Muller). Such systems relieve re-
matical research being stimulated by the use of a search workers of a great deal of drudgery. Indeed,
computer is the soliton (solitary wave) solution of they make possible manipulations which just could
the Korteweg-de Vries equation by Zabusky and not be done manually in any reasonable time or with
Kruskal (1965), which was initially suggested by nu- any valid hope of an accurate result. Another old
merical results. Continuing experimental investiga- topic that has taken on a new lease of life is that
tions have indicated the existence of other, related, of continued fractions, both as providing approxi-
solutions and theoretical research has provided a mations to real numbers and, in analytical form, in
substantial framework for investigating soliton se numerical analysis.
lutions of several nonlinear wave equations. The availability of colour graphics displays and
Another example is found in the work of Ya- packages has opened up exciting possibilities for re-
maguti, which may be summarized briefly by say- search not only in geometry, modelling and fluid
ing that he observed continuous, but nowhere- flow but in less obvious areas such as analysis (see
differentiable, functions via numerical experiments the chapter by Tall and West). The study of the it-
on dynamical systems defined iteratively whose so- eration of complex-valued functions has been trans-
lutions exhibit very chaotic behaviour. Particular formed recently; the complex nature of Julia sets
cases produce the Weierstrass function and the Tak- and their descendants is made beautifully apparent
agi function (see also the chapter by Tall and West); by the use of colour graphics, even through much of
the latter may be written their mathematical nature remains unknown (see,
for example, Section I.4 below).
T(X) = 2 2-&qx) It is clear to us that the computer is having, and
k=l will continue to have, a significant impact on the
where directions of mathematics research and on the way
in which mathematicians carry out their research.
0 < I < l/2 Computers will not only be commonly used to arrive
d(x)={;;l-x) 1/2<I<l.
at conjectures but also to assist in finding proofs.
and has recently been used in teaching elementary In addition, some important questions are raised:
analysis. Further research, in collaboration with (i) How should computers be used to assist math-
Hata on a family of finite difference schemes led to ematicians in communicating their discoveries and
Lebesgues Singular Function. in keeping abreast of the research of others? and
Among long-established branches of Pure Math- (ii) What are likely to be the intellectual and se
ematics where computers have had a major impact cial consequences, so far as mathematics and math-
are Group Theory, Combinatorics and Number The- ematicians are concerned, of the widespread interest
ory. Many applications of computers in these areas in, and use of computers?
have been published in proceedings of conferences
1.2 Proof
(for example, Churchhouse and Herz (1968), Atkin
and Birch (1971), Leech (1970)). In mathematics a proof is, strictly, a chain of
The applications are already too numerous to deductions from the axioms; in practice, of course,
list in full or describe in detail but it is clear that a proof is accepted if it makes use of results which
the search for sporadic groups, the investigation of have themselves been deduced from the axioms, or
Burnsides problem, the study of rational points on from other results, etc., etc. It would be possible,
elliptic curves, and the search for large primes would but exceedingly tedious, to write out a proof of the
be quite impossible without computers. The fac- theorem that every positive integer is the sum of
torisation of large integers is another example; al- the squares of four integers by starting from the ax-
though intrinsically it is not an exciting topic it has ioms of arithmetic, but few people would regard this
recently assumed considerable importance in rela- as necessary and would accept various intermediate
tion to cryptography and public-key systems (Beker steps - an identity of Jacobi, or representation of
14 Inf7uence of Computers and lnformatics on Mathematics and Its Teaching
integers by binary quadratic forms - as valid rungs purely human proofs; many false proofs - includ-
on the ladder, since each of these steps is deducible ing some of the 4-colour theorem - have been pub-
from other results which are deducible . . . . . from the lished in the past; we do not believe that the com-
axioms. puter will increase the number of false proofs, quite
Computers might be used in mathematical the contrary.
proofs; they might, initially, suggest what is true It is, of course, accepted that no amount of nu-
and, equally important, what is not, they might be merical evidence constitutes a proof of a theorem
used for computations which are required in a proof; relating to an infinite set; the numerical evidence
they might be used - as in the proof of the 4-colour may be misleading even for a very large set of val-
theorem (Appel and Haken, 1976) - to examine all of ues of the variables involved. A well-known example
the finite set of cases, on which the truth of the the- from analytic number theory is Littlewoods proof
orem ultimately depends; they might even be pro- (see Ingham, 1932) that despite all the numerical
grammed to find part of the proof by trying many evidence then, and even now, available
possible combinations of known axioms, theorems or
identities, though combinatorial explosion makes
such an approach infeasible except in very special
cases. (where r(x) indicates the number of primes less
As examples, computers have been used to sug- than or equal to x) not only event,ually changes sign,
gest results in group theory, combinatorics, number but does so infinitely often.
theory, coding theory and to support the truth of A criticism of computer-assisted proofs - such as
conjectures such as the Riemann Hypothesis. For an that of the 4-colour theorem - is that they tend to
early survey article see Churchhouse (1973). Among rely on brute-force and give little insight into why
notable theorems which were initially conjectured the theorem is true. Unfortunately some results e.g.
on the basis of numerical evidence are the Prime finding large primes or factoring large integers in-
Number Theorem (Gauss) and several important re- trinsically require such methods, and whilst it may
sults of Ramanujan (1927) including the congruence be true that a computer proof may bring little in-
properties of the partition function and of the func- sight, its very existence may inspire people to find
tion T(R). On the other hand Lander and Parkin more elegant, shorter, or illuminating proofs.
(1967) and a computer found that Taking a longer-term view, the availability of
27 + 84 + 110 + 1335 = 144 computer assistance may encourage mathematicians
to a more precise syntax and to express more for-
and so disproved a conjecture of Euler that had
mally what is in their minds (de Bruijn, 1970). Such
stood for nearly 200 years. One very specific re-
a development may, in turn, aid the teaching of the
cent achievement deserves special mention viz: the
art of constructing proofs and so lead to the develop-
disproof by Elkies (1988) of the Euler Conjecture on
ment of expert systems to undertake a least some
sums of fourth powers. Euler conjectured that (in-
aspects of mathematical work (including all the rou-
ter alia) no fourth power could be the sum of three
tine algebraic manipulation, computation, etc.), in
fourth powers. Elkies however found that
partial fulfillment of Leibnizs dream of a rational
(2682440)4 + ( 15365639)4 + (18796760) calculating device.
= (20615673)4 One final point: Since every proposition that is
and went on to prove that not only are there an in- provable has among its many proofs one of minimal
finity of such counterexamples but, when expressed length and since the proofs of any given length are
as the representation of 1 as the sum of three ra- (at most) finite in number there must be true theo-
tional fourth powers, the solutions are dense in rems of mathematics that cannot be demonstrated
< 0,l >. by traditional discourse within the longest human
Accuracy and reliability of the computations lifetime. It would appear then t,hat there are math-
should not be an issue today. Where a result is ematical theorems that can only be proved with the
sufficiently important or in doubt it can be checked aid of computers if we are unwilling to wait too long.
by someone else on a different machine; this has
1.3 Experimentation in Mathematics
been done on several occasions and if the result is
confirmed and, assuming that the underlying math- Certain branches of mathemat.ics have always
ematics is correct, the result can be accept,ed with been open to experimentation but the arrival of
considerable confidence, if not certainty. Computer- computers means the scope for experimentation in
assisted proofs need not be any more suspect than mathematics has been greatly increased. In some of
the sections above we have indicated cases where ex- be simulated on a computer. Thus we no longer
periments have been used to provide data on which need to place physical models in a wind tunnel but
conjectures and, in some cases, theorems have been instead simulate the model on a computer. Similarly
based. Euler, remarking on the necessity of observa- we do not have to build a new telephone system to
tion in mathematics, said: The problems of num- see if it works since we can first simulate the system
bers that we know have usually been discovered by on a computer.
observation, and discovered well before their valid-
Exploratory Data Analysis
ity has been confirmed by demonstration . . ...
The sheer speed of computers means that calcu- It is sometimes stated that the computer has led
lations which would once have taken a lifetime can to an unwelcome shift from hard thinking to a sense-
now be completed in hours, or even minutes. Add to less computation of examples and experimentation.
this the fact that the results can often, if required, A balanced picture would say that the computer
be presented in graphical form rather than as a list has led to broader variety of types of rationality
of numbers and we see that the interpretation of the to approach problems and it is necessary to judge in
experiments may be made much easier. The case of every situation which approach is more reasonable.
the iteration of complex-valued functions illustrates The classical paradigm for applying statistics is
this point. to think first very hard and then construct a proba-
Of course, when a constraint is relaxed, there is bilistic model and an adequate design for gathering
a danger of excess. The ability to perform calcula- data. But this strategy is not feasible in quite a lot
tions does not mean that everything can or should of situations where little is known about the data
be calculated. There is a balance to be struck and and the underlying system of interest. In connec-
this must be guided by experience - not to mention tion with the numerical and graphical capabilities
the cost of the computations. The effort and cost of computers a new methodology of data analysis,
involved need to be combined with the probability called Exploratory Data Analysis (Tukey, 1977), has
of success, in the sense of solving a problem or un- been developed. The computer has made it possible
covering some useful fact. Computation for the sake to experiment with several models for a data set, to
of computation is not to be encouraged. construct a variety of interesting plots of the data to
Although experimentation in pure mathematics gain insights into patterns, structures and anoma-
has its uses it is, perhaps, in the area of statistics lies of the data and to develop conjectures concern-
that it is particularly valuable. We take two exam- ing the features of the system underlying the data.
ples. Such a type of exploratory mathematics would not
be practicable on a large scale without using com-
Simulation puters.
Even before the availability of the modern
I.4 Iterative methods
computing technology, experimental sampling and
Monte Carlo methods have played a role in statis- Methods of solving systems of linear equations
tics for studying the performance of statistical tech- are traditionally divided into (i) direct and (ii) in-
niques under the assumption of probability models. direct, or iterative, methods. The direct methods
The computer has enhanced this aspect on a large include Gaussian elimination, the indirect methods
scale. One famous example is the Princeton Ro- include the Gauss-Seidel method. Direct methods
bustness Study (Andrews et al, 1972) where sets of have the advantages (a) that they will always pro-
estimators under a system of different modelling as- duce the solution provided that it exists, is unique
sumptions are studied by means of computer simu- and that sufficient accuracy is retained at every
lation. The results have stimulated new mathemati- stage, and (b) that the solution is found after a
cal research into robust estimators (e.g. asymptotic known number of operations. They have the disad-
theory) but on the other hand they cannot merely vantage that very sparse systems of equations, such
be interpreted as conjectures that can and should as arise in finite difference approximations to differ-
be validated by mathematical proof, but they have ential equations, may become rapidly less sparse as
an importance in itself and have already influenced the elimination process proceeds so raising the stor-
the practice of analyzing data. age requirement from a multiple of n (for n equa-
More generally, computers have given a mations) to something like n2. Iterative methods, on
jor impetus to the idea of mathematical modelling the other hand, may fail to converge to a solution
wherein a physical or logical situation is embodied and, if they do converge, it is not obvious how many
in a mathematical model whose operation may then operations they will require to produce the desired
accuracy. They have, however, the very consider- of Barnsley on iterated function schemes, which re-
able advantages that they are very well suited to sult in remarkably lifelike pictures of ferns and trees,
computers and preserve the sparsity of the coeffi- has aroused a lot of interest. The mathematical the-
cient matrix throughout. ory of fractals is much more demanding than their
Direct methods of solution of nonlinear systems production on a computer but good progress has
are rarely available; there is, after all, no direct been made here, undoubtedly inspired by the com-
method for solving the general polynomial of even puter graphics successes.
the fifth degree and so iterative methods are gen-
erally used. As in the case of linear systems, con- 1.5 Algorithms
vergence may not always occur, though conditions An algorithm is simply a procedure for solving
sufficient to ensure convergence are usually known; a specific problem or class of problems. The notion
and although in some cases the number of iterations of an algorithm has been around for over 2000 years
necessary to produce convergence to a specified ac- (e.g. the E UCId 1 ean Algorithm for finding the high-
curacy may not be easily predicted, it is frequently est common factor of two integers), but it has at-
not a matter of great importance and, if time is lim- tracted much greater interest in recent years follow-
ited, accelerating techniques can often be used. ing the introduction of computers and their applica-
The revival of interest in iterative methods tion not only in mathematics but also to problems
brought about by the use of computers has led to arising in technology, automation, business, com-
significant advances in the study of functions which merce, economics, the social sciences, etc. (see also
are iteratively defined, e.g. by a relation of the type the chapter by Maurer). Computer algorithms have
Z n+l = F(Zn) been developed for many commonly occurring types
of problems. In some cases several algorithms have
where Zo is a given complex number and the func- been produced to solve the same problems, e.g. to
tion F(Z) may contain one or more parameters. sort a file of names into alphabetical order or to in-
Some functions of this type, such as vert a matrix, and in such cases people who wish to
use an algorithm will not only want to be sure that
Z n+1 = 2: + c
the algorithm will do what it is supposed to do, but
were studied over 60 years ago by Julia (1918) and also which of the several algorithms available is, in
Fatou (1919), but attracted relatively little interest some sense, the best for their purposes. An algo-
at that time. In the case where the function F(Z) rithm which economizes on processor time may be
involves one complex parameter C and we define the extravagant in its use of storage space or vice-versa
set of points KC to be those points Z such that the and the need to find algorithms which are optimal,
iterated sequence of points given by or at least efficient, with respect to one or more
parameters has led to the development of complex-
Z, F(Z), F(F(Z)), . . .etc.
ity theory. Thus the Fast Fourier Transform has
does not go to 00, then the boundary of Kc is called reduced the time complexity of computing Fourier
the Julia set associated with F(Z) and C. Only re- transforms from order n2 to order n log n, which is
cently, thanks to the availability of computers and, of considerable practical importance for large val-
particularly, of colour graphics terminals has the ex- ues of n. More recently the problem of designing
traordinary nature of these Julia sets and their nu- algorithms which can be efficiently run on several
merous spin-offs been appreciated. For example, the processors working in parallel has attracted consid-
Mandelbrot set is defined as the set of values of C erable interest.
for which Kc is connected. The relation above is a Algorithms which are ideal on a single proces-
fructal curve, the discovery of which, due to Man- sor may be highly inefficient, or even fail entirely,
delbrot, has inspired a great deal of exciting and on parallel processors. The search for algorithms
attractive research by Douady, Hubbard and many for the efficient solution of mathematical problems
others (Devaney, 1989). on systems of parallel computers is a major area of
The enthusiastic study of fractals has grown very research and conferences on this topic are held regu-
rapidly in recent years and the ready availability of larly. The problems are mathematically challenging
high definition computer graphics has made it possi- and are also of considerable practical importance.
ble for schoolchildren, as well as teachers, to produce With parallel computer systems now being readily
a wonderful variety of exotic pictures based upon available, courses on parallel computing are being
iteration of simple functions of complex variables taught at undergraduate level which, five years ago,
(Peitgen et al, 1992). Even more recently the work would have been possible in very few places.
A final point is this: the growing importance of l epistolary - where A writes a letter to B
algorithms suggests an enlarged role for proof by al- communicating the result;
gorithm in which a constrictive proof of an existence l proscriptive - where A writes the result on
theorem is obtained by exhibiting an algorithm to a wall (literal or metaphorical) for others to read;
construct the object posited. l privately published - the usual form is a de-
partmental technical report, whose existence is an-
1.6 Symbolic Mathematical Systems nounced;
The possibility of using a computer to manipu- l publicly published - journals or books.
late symbols, rather than numbers, and so provide This communication may be received either di-
users with packages for algebraic manipulation and rectly by the person who is going to use the result,
indefinite integration was appreciated from the ear- or indirectly.
liest days of computers. Packages such as ALPAK The advent of computed-aided typesetting and
and Slagles SAINT (Slagle, 1963) both date from camera- ready copy has obviously changed the vi-
the early 1960s. Not only were such packages avail- sual form of mathematical comunication (partic-
able, they were used, Around 1960, Lajos Tokacs ularly the publicly-published) and its economics.
used ALPAK to carry out some very tedious alge- This has consequences for mathematicians (espe-
braic manipulation involving 1200 terms to find the cially editors) who may need to read the input to
second moment in a problem in queueing theory, of such type-setting systems. But computer technol-
importance to Bell Laboratories. No one had had ogy is capable of changing and is changing, far more
the courage or energy to do this by hand. When the than this.
second moment was finally found it reduced to just Epistolary. Computer networks have revolu-
three terms, after which a shortened mathematical tionized this method of communication by allow-
derivation was obtained and a general theory de- ing letters to be sent via electronic mail instead
veloped. Two points are worth noting: After the of physical mail As more and more mathematicians
brute- force use of ALPAK the nature of the solu- are linked by such networks, they will replace most
tion inspired mathematicians to find a more elegant written communication.
derivation - in support of our remark in Section 1.2;
Proscriptive. In addition to the physical no-
secondly, without the use of a symbolic manipula-
tice boards in ones own department or elsewhere
tion package it is unlikely that this work would have
on which one can place proofs (or, more likely, an-
been done at all.
nouncements of technical reports containing proofs),
Another early system, FORMAC, was utilized to computer networks distribute electronic bulletin
help with the solution of the restricted case of the 3- boards to various sites which subscribe to them.
Body Problem and, more recently, it has been used In some areas of computer science in North America,
to check that two 752-term polynomials, occurring most results are announced on such bulletin boards.
in the theory of plane partitions, are identical.
Private Publishing. This is closely related to
Some symbolic manipulation packages are gen- the above. Such networks also distribute electronic
eral, but many more are applications specific. We newsletters to individual subscribers, which often
have mentioned CAYLEY which is widely used for contain lengthy articles in draft form, or state con-
the study of finite groups both at research level jectures or problems.
and as a teaching aid. Other specific systems in-
Public Publishing. This is the area whose
clude MATRIX, REDUCE (Fitch, 1985), MAC-
form has been directly least affected. Though there
SYMA (Pavelle and Wang, 1985); Maple (Char,
is talk of it, no serious refereed journals distributed
1988); Mathematics (Wolfram, 1988); many more by electronic means exist.
traditional algebra systems are surveyed in Pavelle
et al (1981). These are further discussed in the chap- REFERENCES
ter by Hodgson and Muller.
Andrews, D.F. et al. [1972]: Robust Estimates and
1.7 Computers and Mathematical Communi- Location, Princeton, NJ: Princeton University
cation Press.
Appel, K. and Haken, W. [1976]: The solution of
Whilst it affords great personal satisfaction to
the 4-color map problem, 51% American (Oc-
prove (or disprove, or conjecture) a result, the math-
tober), 108-121.
ematical community only gains if that result is coni-
municated to others. This communication may take Atkin, A.O.L. and Birch, B.J. (Eds.) [1971]: Com-
various forms (though the distinctions are not rigid). puters in Number Theory, San Diego: Aca-
demic Press. Wolfram, S. [1988]: Mathematics, Reading, MA:

Addison-Wesley.
Char, B.W. et al [1991]: Maple V Library Reference
Manual, New York: Springer-Verlag. Zabusky, N.J. and Kruskal, M.D. [1965]: Interaction
of solitons in a collisionless plasma and the
Churchhouse, R.F. [1973]: Discoveries in number
recurrence of initial states, Phys. Rev. Lett.,
theory aided by computers, Bull. IMA, 9, 15-
15, 240-243.
18.
Churchhouse, R.F. [1980]: Computer arithmetic
and the failure of the associative law, Bull.
IMA, 16, 210-214.
Churchhouse, R.F. [1985]: Computer arithmetic II:
some computational anomalies and their con-
sequences, Bull. IMA, 21, 70-73.
Churchhouse, R.F. and Herz, J.C. (Eds.) [1968]:
Computers in Mathematical Research, Ams-
terdam: North Holland.
Devaney, R.L. [1989]: An Introduction to Chaotic
Dynamic Systems (2nd Ed.) (A-W Studies in
Nonlinearity), Reading, MA: Addison-Wesley.
Elkies, N.D. [1988]: On A4+B4+C4, Math. Comp.,
51, 825-835.
Fatou, P. [1919]: S ur les equations fonctionelles,
Bull. Sot. Math. France, 47, 161-271.
Fitch, J. P. [1985]: REDUCE, J. Symb. Comp., 1,
21 l-227.
Ingham, A.E. [1932]: The Distribution of Prime
Numbers, Cambridge: Cambridge University
Press.
Julia, G. [1918]: Memoire sur literation des fonc-
tions rationelles, J. Math., 8, 47-245.
Lander, L.J. and Parkin, T.R. [1967]: A counter-
example to Eulerss sum of powers conjecture,
Math. Comp. 21, 101-103.
Leech. J. (Ed.) [1970]: Computational Problems in
Abstract Algebra, Oxford: Pergamon Press.
Pavelle, R. and Wang, P.S. [1985]: Macsyma from
F to G, J. Symb. Comp., 1, 69-100.
Peitgen, H-O., Jiirgens, H. and Saupe, D. [1992]:
Fractals for the Classroom. Part One: In-
troduction to Fractals and Chaos, New York:
Springer-Verlag.
Ramanujan, S. [1927]: Collected Papers, Cam-
bridge: Cambridge University Press.
Slagle, J.R. [1963]: A heuristic program that solves
symbolic integration problems in freshman cal-
culus, J. ACM, 10, 507-520.
Tukey, J.W. [1977]: Exploratory Data Analysis,
Reading, MA: Addison-Wesley.
Part II
THE IMPACT OF COMPUTERS AND COMPUTER SCIENCE ON THE

MATHEMATICS CURRICULUM
Anthony Ralston
SUNY at Buffalo, Buffalo, NY 14260, USA
2.0 The Changing Science of Mathematics
In this section we will consider how computers will remain of central importance in the secondary
and computer science should be causing changes in school curriculum. The important thing, however, is
the mathematics curriculum because of the chang- not to have students achieve great manipulative skill
ing importance of various branches of mathematics in algebra (e.g. in polynomial algebra) but rather
wrought by computers and computer science (see to teach them to consider algebra as a natural tool
the chapter by Steen). One aspect of this change for solving problems in many situations. Neverthe-
is that increasingly the knowledge of mathematics less, the ability to use formulas and other algebraic
important to the user of mathematics is no longer expressions will remain necessary.
that of detailed knowledge but rather what might In recent years there has been a trend toward
be called meta-knowledge about the characteris- replacing much of Euclidean plane geometry with
tics and power of methods, often numerical, for the those aspects of geometry more closely akin to al-
solution of classes of problems (see the chapter by gebra. This is useful as a preparation for university
Mascarello and Winkelmann). A related perspec- mathematics but there is much feeling among math-
tive is that computers have brought mathematics ematics educators that the loss of Euclidean geome-
much closer in philosophy to the classical natural try is a sad development. A consensus on how geom-
sciences where there has always been an interplay etry might best be taught at school and university
between theory and experiment. Now mathemat- is not yet available. It should be noted, however,
ics, too, has a laboratory - the computer - on which that some computer scientists feel that the aspect of
experiments can be performed which lead to the- traditional instruction in geometry concerned with
ories and on which theories can be tested. These teaching the meaning and construction of rigorous
points should be kept in mind in what follows. Al- proofs can be achieved through material concerned
though we shall not return explicitly to them, they with the analysis and verification of algorithms (see
influence much of this section. the chapter by Maurer).
2.1 The Common Mathematical Needs of For many parts of mathematics trigonometry is
Students in Mathematics, Science and Engi- useful preparation. But we note that much of the
neering tedious work which was necessary in the past, both
numerical and symbolic, can now be easily done on
hand-held computers.
(a) Preparation for University Next we mention calculus. In many countries
Mathematics
this has been a secondary school subject for many
To provide a context in which to discuss the im- years for most university-bound students while in
pact of computers and computer science on curricu- other countries only the very best students begin
lum and pedagogy, it is necessary to agree first, in calculus in secondary school. The main thrust of
general, on the appropriate mathematics for the sec- secondary school calculus has been to provide stu-
ondary school student (see the chapter by Graf et dents with techniques, and to prepare those intend-
al) and then to consider the university curriculum. ing to study mathematics at university with a first
Since there are significant differences between dif- introduction to the concepts they will encounter at
ferent parts of the world on when secondary school the university level. Since all the techniques of sec-
ends and university instruction begins, the com- ondary school calculus as well as much of univer-
ments which follow will have to be interpreted in sity calculus can now be done on hand-held devices
the local context. or on symbolic mathematical systems (often called
Algebra has traditionally been an important computer algebra systems) on computers (see the
subject in high school. Since elements of abst,ract chapter by Hodgson and Muller), calculus a.t the sec-
algebra are likely to become increasingly important ondary as well as university level must focus almost
in mathematics education, it is clear that algebra entirely on concepts and not on computation.
19
Various new subjects have become part of the social and management science students, might be
secondary school curriculum in recent years. Among about equal for mathematics students themselves
these is probability which has come into the cur- and surely should be weighted more toward tradi-
riculum in many countries. Topics such as discrete tional continuous mathematics for physical science
probability distributions, the binomial distribution and engineering students. It needs to be empha-
and related topics are appropriate. So too, is an sized, however, that all groups of students need
introduction to data analysis and elementary statis- some exposure to both the continuous and discrete
tics because of their importance in science as well as approaches to mathematics. Whether students are
mathematics. Another subject, about which there exposed to calculus first and then discrete math-
will be further discussion below, which we would ematics or vice versa will depend on the student
like to see more of in the secondary school curricu- population and on institutional convenience.
lum, is discrete mathematics including elementary The actual content of the discrete mathemat-
combinatorics and graph theory as well as an intro- ics component is still quite variable. However, the
duction to induction and recursion. In this connec- discrete component normally contains at least some
tion it would be appropriate to introduce both the traditional discrete mathematics (e.g. combina-
design and verification of a number of important al- torics, graph theory, discrete probability, difference
gorithms such as those for sorting. Finally we note equations) as well as perhaps some abstract algebra
that elementary linear algebra, particularly matrix although the latter may follow in a later course after
algebra and work with systems of linear equations, completion of the core courses.
should certainly be considered for the secondary We note also the importance of mathematical
school curriculum. logic in the core university curriculum. Although
(b) The University Mathematics traditionally an advanced undergraduate or a post-
Curriculum graduate subject (at which levels there will be a
The core of the university mathematics curricu- continuing need for specialized courses), logic is so
lum for many years has been the calculus and, to a important in informatics that it needs to be intro-
lesser extent, linear algebra. This is the case no mat- duced early in the university mathematics curricu-
ter how much mathematics the student may have lum. Moreover, with the increasing need for people
studied in secondary school. Computers themselves in the scientific and technical professions to han-
have an impact on both the content of this curricu- dle information in a precise manner, logic has great
lum and its pedagogy. Not only do computers allow value for a wide variety of students. Logic is an im-
more interesting and effective presentation of clas- portant constituent of many discrete mathematics
sical subject matter but, in addition, as with the courses (see below). But it can also be considered
secondary curriculum, they affect what subject mat- as a subject for a course by itself which would fol-
ter is important to students. For example, symbolic low the introduction in discrete mathematics. Such
mathematical systems suggest a deemphasis on the a course can usefully be given a distinctive computer
more skill-oriented portions of the current curricu- flavor as described by Schagrin et al (1985).
lum. As a final matter, we stress the importance of
Informatics (i.e. computer science) itself also using the paradigms of informatics (e.g. an algo-
implies changes in the content of the core curricu- rithmic approach, iteration, recursion) in the teach-
lum. This is essentially because informatics is a ing of mathematics at all levels. Although these
highly mathematical discipline whose problems re- paradigms may seem most easily applicable to dis-
quire almost universally the tools of discrete rather crete mathematics, there is considerable scope for
than continuous mathematics. Thus, there is now their introduction into the classical continuous cur-
a strong argument to provide a balance in the riculum.
core curriculum between the traditional continuous The reader may be surprised to find no mention
mathematics topics and topics in discrete math- of numerical analysis here (or hereafter in this doc-
ematics (Ralston 1981, Ralston and Young 1983, ument) because this subject is the one that most
Ralston 1989). For university courses aimed at a obviously combines the continuous and discrete ap-
broad spectrum of mathematics, science and en- proaches to mathematics. But we take the posi-
gineering students, this balance may well contain tion that numerical analysis is now such a well-
nearly equal portions of the continuous and the dis- established subject in the mathematics curriculum
crete. For those courses aimed at specific student that it does not need to be discussed in the context
populations, the balance might be weighted more of this report. This is, however, not to say that the
in the direction of the discrete for informatics and subject mat.ter of numerical analysis is no longer af-
Impact of Computers on the Mathematics Curriculum 21
fected by advances in computing; developments in, gorithms and an introduction to the predicate
for example, parallel computing are having great im- calculus.
pact on numerical analysis. 8. Infinite processes in discrete mathematics: Se-
quences, series, generating functions, approxi-
2.2 A Discussion of Particular Curriculum mation algorithms.
Areas on Which Computers and Informatics In addition, other possible topics depending
Have an Impact upon local needs and desires are:
Although discrete mathematics and calculus are 9. Algorithmic linear algebra including the use of
discussed separately in what follows, it should be Gaussian elimination as an entree to abstract
emphasized that there is no intellectual reason to linear algebra and an introduction to linear pro-
consider them as separate subjects. Indeed, they are gramming and applications of linear algebra.
mutually supportive and ideally would be taught in 10. Decision mathematics including such things as
integrated courses (see the chapter by Seidman and queueing theory and packing problems.
Rice). However, for at least some years to come,
11. Algebraic structures such as rings, groups etc.
such integrated courses will be relatively rare, not
12. Finite state machines and their relation to lan-
least because of the lack of textbooks for integrated
guages and algorithms.
courses.
And, of course, there can be extensions of all the
(a) Discrete Mathematics Courses above topic areas to more advanced subject matter
We begin with a discussion of what topics in dis- if desired and appropriate.
crete mathematics should be contained in courses Since the Strasbourg conference in 1985, at least
intended for mathematics students as well as for 40 books have been published from which a course
students in the social and management sciences. Al- on the above lines can be taught (see, for example,
though the topics to be listed below cover a broad Epp, 1990 and Maurer and Ralston, 1991).
spectrum, it is possible to design a coherent course The experience of those who have taught such
covering these topics if the course is built around courses is that, despite the potpourri of topics listed
themes such as algorithms and their analysis and above, these courses can be made interesting and
inductive and recursive thinking. satisfying if a consistent, coherent approach is taken
A Discrete Mathematics Syllabus which emphasizes algorithmic, recursive and induc-
tive thinking.
1. Mathematical Preliminaries - Sets, functions, re- Following a course from a syllabus like that
lations, summation and product notation, ma- above, a variety of advanced courses in discrete
trix algebra, an introduction to proof and logic mathematics can be contemplated although only
concepts. the largest institutions would be able to offer all
2. Mathematical induction including its applica- of these. Indeed, each of the subject areas listed
tion to algorithms and recursive definitions. above suggests one or more advanced courses which
3. Graphs, digraphs and trees including path, would build on the introductory material in a first
searching and coloring algorithms, tree traversal, discrete mathematics course. Most of these courses
game trees and spanning trees and applications are currently in a process of evolution as the sub-
in a variety of areas. ject matter in the first discrete mathematics course
4. Basic Combinatorics including the sum and changes and develops and as the applications of dis-
product rules, permutations, combinations and crete mathematics grow and diversify. A program
binomial coefficients, inclusion-exclusion, the which combines a carefully constructed introductory
pigeonhole principle and combinatorial algo- discrete mathematics course with several advanced
rithms. courses will give the student a firm basis for study-
5. Difference equations (i.e. recurrence relations) ing informatics as well as providing a basis for pro-
including first order equations, constant coeffi- fessional work in modern applied mathematics and
cient equations and the relationship of recur- other fields in science and engineering.
rence relations to the analysis of algorithms, par-
ticularly divide-and-conquer algorithms.
6. Discrete probability including random variables,
discrete distributions and expected value.
7. Mathematical logic including the propositional
calculus, Boolean algebra, the verification of al-
22 Influence of Computers and Informatics on Mathematics and its Teaching
(b) Calculus in the Computer Age clude symbolic mathematical systems and a variety
(i) The Role and Relevance of Calculus of graphical packages. Advances have taken place
so rapidly in these areas that it is now the case that
Among the key factors which compel change in very powerful symbolic and graphical systems are
the teaching of university mathematics courses are: available on hand-held computers (e.g. the BP-48s)
- the substantial experience with minicomput- as well as on microcomputers. One result of this is
ers and microcomputers and programming packages that an understanding of functions, variables, pa-
which many students have had before coming to the rameters, derivatives etc. and the ability to inter-
university; pret formulas and graphics is becoming more im-
- the growth of new areas of applied mathemat- portant to the student than skills in executing the
ics such as the analysis of algorithms and computa- (numerical or symbolic) procedures of calculus. In
tional complexity. the teaching of calculus to all students the need is
One result of this is that many students have at- clear for a shift from an emphasis on calculational
titudes and expectations which lead them to believe technique to one which emphasizes the development
that the most challenging and meaningful mathe- of mathematical insight.
matical problems today are related to computers (ii) The Content of Calculus Courses
and informatics. This cannot help but influence
how we must motivate mathematics students and If functional behaviour and representation are
all other students in mathematics courses. to be the focus of the calculus course, then continu-
In considering the place of calculus in the com- ous functions and discrete functions (i.e. sequences)
puter age, we cannot forget that it is one of hu- must be emphasized and motivated by a wide vari-
mankinds great intellectual achievements. Every ety of mathematical models. (Indeed, mathemat-
educated person should be aware of it. Its his- ical models and their applications in a variety of
tory exemplifies the unreasonable effectiveness of disciplines should be an important part of calcu-
mathematics better than any other branch of math- lus courses.) (Note: it can be argued that se-
ematics. And its effectiveness is as great today as quences belong more properly in the discrete math-
it has ever been. But this does not excuse teaching ematics course discussed previously. This only illus-
calculus as is so often the case now with an empha- trates the need to bring the discrete and continuous
sis only on the execution of mechanical procedures points of view together into an integrated sequence
- and paper-and-pencil procedures at that. Instead of courses as soon as possible.
calculus needs to be taught to illustrate the unique An important theme in calculus courses should
ways of thinking it epitomizes. be the contrast between the local and global be-
The realm of applications of calculus remains im- haviour of functions. Local behaviour is, of course,
mense. They are, indeed, increasing due to the in- derived by studying the derivative for continuous
creasing mathematization of heretofore qualitative functions (and the difference operator for discrete
sciences like biology. In constructing calculus mod- functions). And similarly the integral (and summa-
els of phenomena and then solving the resulting tion) operators are used to derive global information
equations, there is often an interplay between these about functions. Undoubtedly it will remain neces-
models and their discrete counterparts with the cal- sary to develop some ability to do formal computa-
culus models representing the limiting behaviour of tions with derivatives and integrals. But the major
the discrete models. It is now more important than emphasis should be on numerical algorithms (par-
ever to include this interplay in calculus (and dis- ticularly for integrals) and on how derivatives and
crete mathematics) courses because inevitably the integrals can be used to understand the behaviour
solution of most problems in calculus involves the of functions.
(discrete) computer. The discretization necessary A topic such as the Taylor series representation
to solve problems of calculus with a computer often of a function should be used to show how good local
has not borne a close relationship to the underlying information can be obtained using low-degree Taylor
discrete model. But the increasing power of com- polynomials and interpolating polynomials, another
puters means that more and more frequently it is area where the analogy between the continuous and
possible to have computer models which mirror very the discrete may be usefully shown.
closely the discrete models from which the continu- Finally, there should be a balance in the cal-
ous model was initially abstracted. culus course between traditional topics and ones
There already are powerful software tools which whose importance has greatly increased because of
can be used in the study of calculus. These in- the advent of computers and informatics. Thus,
Impact of Computers on the Mathematics Curriculum 23
for example, the O() and 00 notations, which en- fundamental concepts and theorems.
able the asymptotic growth rates of functions to To take full advantage of the use of computers in
be compared to standard functions like polynomials teaching calculus, it will be necessary to change the
and logarithms, are not always taught in calculus standard classroom environment. Classrooms need
courses, but they should become so. to be provided with large monitors or screens on
This discussion is intended only to provide the which the monitors may be projected. Both inside
flavor of how an orientation toward computation and outside the classroom, students need adminis-
should change the approach toward teaching most tratively easy and user-friendly access to comput-
of the standard calculus topics. ers and software. Teachers will need private com-
puter facilities in order to prepare course material.
(iii) Computers for Learning and
A prerequisite for this is in-service training so that
Teaching Calculus teachers may become comfortable with computers
Computers enable teachers to modify their and then fluent in their use and aware of possibili-
methods of teaching calculus (and, of course, much ties beyond what may be available in the particular
other mathematics also) in order to meet better the software on which they have learned.
need of their students. Computer graphics is a pow- Finally, we note the value of using computers
erful medium in which to provide examples - and in the classroom to teach mathematics. The desir-
non-examples - of continuous functions, discontin- ability of this for calculus and related subjects is
uous functions, the area under a curve, direction particularly clear since the dynamics of computer
fields and nowhere differentiable functions as well as graphics is ideally suited to help explain a subject
in many other areas. Well-designed software (there which is essentially about change. Indeed, it is ironic
still isnt nearly enough of this) can be used by that only static technology - the chalkboard and the
students to discover and explore the concepts men- overhead projector - are still used so widely to teach
tioned above as well as such fundamental concepts calculus. There is a considerable amount of software
as slope and tangency (see also Section 2.3). But the available now which can be used in the classroom to
effective use of such software requires that teachers teach calculus (e.g. Flanders (1991)) and differen-
sometimes depart from a lecturing style and go in- tial equations. There is much less software available
stead to guiding and interacting style with small to teach discrete mathematics in the classroom but
groups of students or individual students. there are numerous aspects of discrete mathemat-
Well-designed software will also permit enhance- ics (e.g. induction and recursion) for which suitable
ments by students through the writing of (usually software would be valuable in the classroom. We
short) programs. This is just another way in which can expect to see the development of such software
students can be actively involved in their own learn- in the near future.
ing although it is important that the use of the
2.3 Exploration and Discovery in
computer does not become the message instead of
Mathematics
the mathematics which it is supposed to illustrate.
Thus, programming per se should not play any sig-
nificant role in a calculus course. The idea of using computers to enable students
Another impact of the computer in calculus may to explore mathematics and discover mathematical
be to change the order in which topics are taught. patterns for themselves is not a new idea (Steen,
For example, it is becoming increasingly common 1988). However, the advent of powerful and avail-
to introduce limits at the very start of a calculus able computer systems makes this point so impor-
course. Tangent functions and area under a curve tant in teaching mathematics today that we devote
can be motivated and defined graphically. When an entire section to it.
a formal definition of a limit is needed, students First, why should exploration and discovery be
will be ready for it. As another example, differ- important components of the educational process in
ential equations can now be treated much earlier mathematics? The answers parallel the reasons why
in the curriculum than was previously possible be- we teach mathematics in the first place:
cause of the ease of understanding made possible by l active learning leads to better retention and un-
new graphics systems (see the chapter by Tall and derstanding and more liking of the mathemat-
West). They can be introduced right after differ- ics we teach because the mathematics is seen
entiation and before integration. Studies are now as a basic component of human culture; it also
under way to discover whether such reorderings will leads to more self-confidence in the ability to use
lead to a greater or more rapid understanding of mathematics to solve problems;
24 Inffuence of Computers and lnformatics on Mathematics and Its Teaching
exploration and discovery helps to teach people can be overcome; teachers can be trained to feel
to think; comfortable with this mode of learning.
discovery provides the greatest aesthetic expe-
2.4 Some Speculation about the Future
rience in mathematics, the aha of seeing or
proving something is what makes mathematics As mathematics becomes increasingly an exper-
attractive; imental science, it is inevitable that computers and
exploration and discovery are perhaps the best computer science will have increasing influence on
ways for students to see that mathematics is so the mathematics curriculum. Computer science will
useful; become a gradually greater focus of applications of
mathematics and this will affect what is important
discovery enables the student to see a familiar
in mathematics. At the same time the means by
idea applicable in a new context, thereby en-
which all mathematics is taught will be inextrica-
abling a grasp of the power and universality of
bly entwined with computer technology. Although
mathematics.
the cost of this technology will continue to be a
Computer technology may be used to assist in
problem for developing countries, the curricular in-
mathematical exploration and discovery in a variety
ertia in developing countries is far less than that
of ways; for example:
in the developed countries. Developing countries
through visualization of a great variety of two have an unparalleled opportunity to use computers
and three dimensional objects via computer and the influence of computer science to modernize
graphics, students may explore questions and their mathematics curricula and their mathematics
discover results by themselves. teaching faster than will be possible in developed
through computer graphical presentations of in- countries.
teresting geometries like flatland and turtle
geometry; REFERENCES
via exploratory data analysis to, for example, Douglas, R.G. (Ed.) [1986]: Toward a Lean
draw conclusions from data (e.g. is it bimodal? and Lively Calculus, MAA Notes Number 6,
are there outliers?), to transform data (e.g. by Washington, DC: Mathematical Association of
logarithmic plots), to smooth data and to com- America.
pare different sets of data. Epp, S.S. [1990]: Dzscrete Mathematics with Appli-
by graphical and numerical explorations of how cations, Belmont, CA: Wadsworth.
to approximate complicated functions by simple
Flanders, H. [1991]: Microcalc 6.0, Calculus Soft-
ones;
ware for VGA, Ann Arbor, MI.
by applying the first step of the inductive
paradigm - compute, conjecture, prove - in Maurer, S.B. and Ralston, A. [1991]: Discrete Algo-
many, many different situations; rithmic Mathematics, Reading, MA: Addison-
by using symbolic mathematical systems to dis- Wesley.
cover mathematical formulas such as the bino- Ralston, A. [1981]: Computer science, mathemat-
mial theorem; ics and the undergraduate curricula in both,
by designing and executing different algorithms Amer. Math. Monthly, 88, 472-485.
for the same or related tasks. Ralston, A. (Ed.) [1989]: Discrete mathematics in
This list could be made much longer. Readers will the First Two Years, MAA Notes Number 15,
probably be led to make their own suggestions. Washington, DC: Mathematical Association of
There are various implications to using comput- America.
ers to facilitate exploration and discovery: Ralston, A. and Young, G.S. [1983]: The Future
l we must start with easy tasks so that students of CollegeMathematics, New York: Springer-
feel they are really succeeding on their own and Verlag.
are not being led step by step by the teacher;
Schagrin, M.L., Rapaport, W.J. and Dipert, R.R.
l teachers need to be educated for this kind of in-
[1985]: Logic: A Computer Approach, New
structional mode; few teachers can become com-
York: McGraw-Hill.
fortable with these ideas without explicit educa-
tion; we note, in particular, that testing what Steen, L.A. [1988]: The Science of Patterns, Science,
has been learned by the student is not easy. But 240 (29 April), 611-616.
experience has shown that success is not only
possible but yields rich rewards. The difficulties
Part III
COMPUTERS AS AN AID TO TEACHING AND LEARNING MATHEMATICS
B. Cornu
IUFM, Grenoble, France
3.0 Introduction
Mathematicians and mathematics teachers have to effect major changes in teaching and learning in
been provided with a new tool, the computer. There the future.
is no shortage of applications or interesting exam- Computers for mathematics teaching are not so
ples which one can quote. But, like all tools, the widely used as one could think. Appropriate soft-
computer by itself does not supply a solution to ware and strategies do exist; but they are used
the problems of mathematics education. There is by few teachers; one of the main problems now
no automatic beneficial effect linked to a computer: is to help all teachers to use computers, not as a
The mere provision of micros in a class - or lecture new experience, but as a common tool for teach-
room will not solve teaching problems. It is essen- ing. This requires not only good training for teach-
tial, therefore, that we should develop a serious pro- ers and good pedagogical products and tools, but
gramme of research, experimentation and reflective also good integration of new technologies in curric-
criticism into the use of informatics and the com- ula and good long term pedagogical strategies.
puter as an aid to teaching mathematics. It will not
suffice to think only in terms of mathematics and 3.1 A changing view of mathematics
the computer, and of the production of software
which amuses and interests mathematicians. We There are many references in this book to the
must also take into account types of knowledge and way in which the computer can lead to a changed
the ways in which these can be transmitted, and at- view of what mathematics and mathematical activ-
tempt to study, in a serious epistemologically-based ities comprise. For example, as the experimental
manner, various concepts and the obstacles which aspects of mathematics assume greater prominence
they present to learners. We must think of students, (see Section 2.3), and there is a corresponding wish
their development and the matching of new and old to ensure that provision should be made for students
knowledge. We must consider in depth the teaching to acquire skills in, and experience of, observing,
possibilities created by the computer. It is essential, exploring, forming insights and intuitions, making
above all, that we should move beyond the stage of predictions, testing hypotheses, conducting trials,
opinions, enthusiasms, and wishful thinking and en- controlling variables, simulating, etc. Examples of
gage in a true analysis of the issues. Only in this way how such work can be carried out are found in later
will we come to a true resolution of certain problems chapters in this book. However, mechanisms need to
of teaching. Such research, of necessity experimen- be found for disseminating information about fruit-
tal, will have to be critically evaluated. It must be ful experimental environments and how these can
shown how, in given circumstances, the use of the be formed.
computer can facilitate the acquisition of a partic- Yet, as we put new emphasis on the particular
ular concept. Finally, such research work will have activities listed above, it is also necessary to ensure
to be built upon and developed to provide a vital that such traditional activities as proving, general-
component in the training (whether formal or self- ising and abstracting are not neglected or omitted.
directed) of teachers and lecturers. Only then can We will need to find an appropriate balance between
computers have any large-scale effect on mathemat- experimental and more formal mathematics.
ics teaching. The possibilities presented by the computer will
Certainly such research has been done in the actually help focus our attention on the kind and
past few years, and we can now see examples of uses types of knowledge which we wish students to ac-
of computers in education, based on a serious st,udy quire. Not only are new possibilities offered to us,
of the didactical problems to be solved. In such but also a greater incentive to ident,ify more pre-
uses, the computer is not a tool supplementary to cisely our educational goals.
traditional teaching; it is integrated in a pedagog- If our aims of teaching change significantly so
ical strategy, adapted to the actual obstacles the as to encompass and stress the process of mathe-
students have in learning. But much remains to be matics more than the products of the mathemat-
done. Both the development of educational research ical activities of others, then there will, of course,
and the evolution of technology have the potential be a need to identify those parts of mathemat,ics
25
26 InA uence of Computers and Informatics on Mathematics and Its Teaching
most suitable for our purposes. Topics and areas of This expression, which troubles so many students,
mathematics must be selected which encourage and can then be thought of in terms of a simple, famil-
facilitate an experimental approach. iar and useful computer process. Again, in Pascal,
Finally, in this section we must stress two impor- representing a fraction as a record with two inte-
tant, interrelated points. Many, indeed the major- ger fields (the second being non-zero) helps students
ity, of our students do not intend to become math- think about rational numbers as ordered pairs of
ematicians. We must not lose sight of the implica- integers, especially if they are given the experience
tions of this in terms of educational goals and em- of writing programs to implement the arithmetic of
phases. But, also, many of these may be students fractions without truncation.
of the experimental sciences. This raises further im- More generally, many mathematical concepts
portant issues, for experiments in mathematics dif- can be defined or described as procedures. This
fer somewhat from those in the physical and natu- gives a more dynamic approach, and can help the
ral sciences. The techniques are often very similar, student in understanding and in using these con-
but in mathematics we have that extra, vital ingre- cepts. Algorithmics (see the chanpter by Maurer)
dient of proof. Experiments are an essential and gives many tools for introducing mathematical con-
neglected part of mathematics, yet mathematics is cepts in such a way.
not an experimental science. The distinctions be- Many examples of ways in which such ex-
tween disciplines and ways of thought will have to periences can be incorporated into mainstream,
be displayed and observed. tertiary-level courses are available. Moreover, the
success of such initiatives would seem to be inde-
3.2 Computers change the relation between pendent of several issues which in discussion tend
teacher and student to be overrated. An important factor in this ap-
Computers can affect the behaviour of students. proach appears to be that students should write the
This creates new interactions and relationships be- programs and so must be cognitively active about
tween student, knowledge, computer and teacher. the processes and data structures they are imple-
The role of the teacher in such situations demands menting. These experiences are then coordinated
considerable thought. with classroom activity.
In their chapter, Mascarello and Winkelmann
(a) The mathematical activity of the stu-
describe a course containing continuous topics such
dent
as multiple integration and ordinary differential
Students will be better able to learn conceptual equations. Here the students wrote programs in
material and develop autonomous (as opposed to a low-level language running on a microcomputer.
imitative) behaviour patterns with respect to math- These were interactive and the results were used for
ematical ideas if they can be cognitively active in experimentation and demonstration.
response to mathematical phenomena presented to Of course, writing programs is not the only use-
them. This activity should consist of the formation ful way in which students can use the computer.
of mental images to represent mathematical objects The use of complicated software packages for il-
and processes. It should also include the develop- lustration of phenomena that are very difficult to
ment of skills in manipulating these objects and pro- display otherwise can clearly broaden the students
cesses. In this way students can increase their abil- awareness and add to their general understanding
ity to think mathematically. (see, for example, the chapter by Tall and West).
Inducing students to emerge from passivity and They can, of course, also be used for exploration
to think actively about mathematics is, however, and discovery. Indeed, some would see the most
not easy. One approach is to make use of the com- exciting opportunity offered by the computer to be
puter to supply sufficiently powerful and novel expe- the way in which it can motivate students to exer-
riences to stimulate such behaviour. The action of a cise the process of discovery. Here we should only
computer program and the structure of data as it is stress the need to see exploration and discovery as
represented in the computer can form useful mod- essential mathematical activities to be practised.
els for thinking about mathematical entities. For Traditionally, this has not been so - teaching and
example, a WHILE loop whose body is a simple learning have been almost wholly concerned with
sum is a process that can represent the mathemati- the transmission and reception of accepted mathe-
cal entity matical facts. However, now, for example, computer
gxi. symbolic mathematical systems (see the chapter by
i=l Hodgson and Muller) permit such rapid and flaw-
Computers in the Teaching and Learning of Mathematics 27
less processing of non- trivial examples that it is at a time when teachers will still be fighting hard
easy first to look for patterns which suggest con- to gain new skills and acquire confidence in them.
jectures and generalisations, and then to search for It would be foolish to underestimate the challenge
counterexamples or machine-aided proofs. this presents.
Computers then can greatly assist us in extend- The acquisition of new skills will be time-
ing the range and the depth of students mathemat- consuming and constantly changing hardware and
ical activities. In some approaches the students will software will make the process a continuing one. For
write their own programs (and there will be an at- many mathematicians these new skills will be read-
tendant risk that mathematical aims may become ily usable in their research work. Others may be
obscured by some of the programming problems); tempted - particularly when universities and other
in others students will use prepared software. Both educational institutions are under pressure - to feel
approaches have already been shown to be of great that such time would be more profitably spent in
. value; further investigations into both will now have increasing personal research output, rather than
to be continued. in improving their teaching, particularly if this re-
quires such a large step in the dark.
(b) The role of the teacher
Computer usage is still actively avoided by many
The computer can be used in two distinct ways mathematicians and by many mathematics teach-
in the classroom. In one it is an aid for the teacher, ers. The problem at the tertiary level is particu-
an electronic blackboard - more powerful than the larly great, for the gulf between the traditional lec-
traditional blackboard, the overhead projector, or a ture often given to a hundred or more students and
calculating machine - but nevertheless a tool whose the classroom/laboratory in which students interact
output is almost entirely under the teachers con- with computers is enormous. To bridge this gulf will
trol. In this role the computer does not upset the need considerable investment in both material and
traditional balance in the classroom. It will still human resources. Time, assistance and in-service
demand effort on the teachers part to select or pro- training will have to be provided on a scale un-
vide suitable software and it can give rise to ex- precedented at this level. Particular attention will
tra administrative problems; in return it should en- have to be directed at those teachers who still have
hance learning. However, it will not revolutionise many years - even decades - to go before they re-
the classroom. tire from teaching. First, however, the necessity for
If, however, students are allowed and expected to change will have to be accepted, and this will only
interact with computers, then the position changes, come through clear, unequivocal demonstration of
for this leads of necessity to a change of method- the benefits which can accrue from innovation.
ology. The teacher no longer has total control - The current problem now is to make all teach-
his/her role can no longer be limited to exposi- ers able to use computers in teaching, or to know
tion, task-setting and marking. The format lecture- why they will not use them! This leads to different
examples, homework-exam must be augmented by, problems:
for example, project (through interaction between l The availability of computers in the teachers en-
student, machine and teacher) assessment on the vironment: Can they easily find and use a com-
basis of a completed (and possibly debugged) as- puter at home for preparing their teaching and
signment. elaborating activities? Can they easily find and
Probably the teacher must combine diverse uses use a computer in the school? Are computers
of the computer. Some activities fit well with the easily available in classrooms?
blackboard computer; some others will be more l The user-friendliness of hardware and software:
efficient if each student has the opportunity to in- Will it take hours and hours to prepare a lesson
teract with the computer. with computers, and will very specialised abili-
Such a change would produce a revolution in ties to use such software be needed?
most class- and lecture-rooms. It demands that l The integration of the computer in the teach-
teachers should not only acquire new knowledge, ing strategy and in the learning environment.
skills and confidence in the use of hardware and The computer is a tool among others, and its
software, but that they should also radically change use must be integrated in a pedagogical strat-
their present aims and emphases, and accept a less- egy. Text,books, homework and all the activities
ening in the degree of control which they presently of the learner must take this into account.
exert over what happens in their classrooms. This The computer does not only change the teachers
last demand means a sacrifice of traditional security, role, but also the attitude and activities of the
students. The relationship is not only between figures very easily, and to modify them by moving
the teacher and the learner: The computer takes some elements (points, lines, etc.), and see at the
its place in the relation, and it also develops the same time how the figure evolves. Invariants and
group work and the project activities. Learning loci can be visualised in a very user-friendly way.
from pupil-pupil talk is one of the components of Such software can be used by the teacher for demon-
the learning environments provided by new tech- strating or by the students in an interactive way.
nologies. Different types of environments, differ-
(b) Many types of utilities are available for use
ent teaching methods and different strategies can
in teaching. Spreadsheets are the best known ex-
be used by teachers. Computers change the organ-
ample. They provide a good environment for intro-
isation of education, and give teachers the role of a
ducing many concepts in arithmetic, algebra, and
pedagogical-engineer in elaborating their strategy,
even calculus. At a very elementary level, they per-
in preparing their teaching, in choosing among the
mit interesting activities about the concepts of vari-
resources available and the tools and products they
ables, parameters, unknowns, etc. They also pro-
will use.
vide nice illustrations of iteration. They are increas-
ingly used in teaching.
3.3 Some particular uses of the computer in
the classroom (c) Databases are now more easily accessible.
They suggest documentation activities, they allow
We have already remarked on the way in which students to look for sophisticated information and
computers can assist in the introduction, develop- so develop project work. They also give teachers the
ment and reinforcement of mathematical concepts, possibility to use or build large sets of exercises and
in building up intuition and insight, etc. In this sec- activities. The distant interrogation of databases is
tion we look at particular ways in which they can now common and enlarges the resources for teach-
be used within the classroom. ers.
(a) Graphic possibilities (d) Artificial intelligence and problem-solv-
ing tools are developing. The first step is to have
Many of the applications of computers in teach-
software and tools able to solve mathematical prob-
ing make use of the possibilities provided for graphic lems. The second step is to produce software able
display. There can be no doubt about their value in
to help students in problem solving.
providing quickly good quality graphic illustrations
which can help build intuition. The example of or- (e) Hypermedia and multimedia products:
dinary differential equations such as z = z2 - t, These allow the integration of different media, and
whose solutions cannot be written down in elemen- their combination for educational uses. They al-
tary terms, is now widely known and used: Visu- low activities which are not linear, but in which
alising the field of tangents and visualising many users may build their own paths and organise their
solutions of the equation make the student better own learning. They considerably enrich other ed-
able to understand the concepts which intervene in ucational tools, linking and making simultaneously
this domain. Moreover, this allows them to discuss available all existing types of software and other ed-
exciting questions concerning the behaviour of solu- ucational technology products. We surmise that in
tions. the future this domain will provoke great changes in
Where the computer scores over many other me- the use of technology in education.
dia is that graphics capabilities now enable move-
(f ) Self-evaluation and individualised in-
ment as well as static diagrams to be portrayed.
struction
This, of course, was true of the film. Yet now the
possibility of being able to change parameters adds a One of the advantages of the computer is that it
completely new dimension to the teaching/learning helps the individualisation of teaching and learning.
experience. This is not only because the student can sometimes
Much interesting and high quality graphic soft- work alone with the computer, but mostly because
ware is now available and allows visual representa- the computer can help to provide a teaching envi-
tions from areas such as calculus, differential equa- ronment which matches the needs of each student
tions, linear algebra, numerical analysis, and geom- ~ the way he learns, the right speed for her, the
etry. appropriate activities.
A famous example in geometry is that of Cabri- The computer can provide a tool for self-
GeometrC which allows pupils to draw geometrical evaluation and can help students to take charge of
the organisation of their own work. It is a diffi- (h) Pocket calculators must be mentionned
cult problem for students to judge how well they here. Even if their possibilities are small in com-
are coping with a subject. One use of comput- parison with computers, they are improving very
ers is to enable students to test themselves. Ques- rapidly: We now have calculators with graphic
tion banks can be made available and instantaneous possibilities and even with symbolic capabilities.
scores given. And the permanent availability of pocket calcula-
The advantages of Computer-assisted Learning tors gives them great power. In many countries, the
(CAL) for individualised instruction have, of course, use of pocket calculators in mathematics has been
been argued for some twenty years, namely that the introduced into the curricula so that all teachers and
computer can offer non-threatening, individualised all pupils use calculators.
responses to students. There have, indeed, been sev- (i) Student errors
eral demonstrations of the value of CAL, for exam-
ple, PLATO in the USA. However, as the cognitive Related to the possibilities described above is
complexity of what has to be learned increases, the that of investigating the errors which students make
difficulties of producing adequate software become in learning mathematics. Such information can be
very great. used in two ways: To help the student remove mis-
The problems become less pronounced when the conceptions, which is its role in individualised CAL,
aim of the program is to revise or to exercise and not or to help the mathematics educator to identify spe-
to teach. Thus Recalling Algebra and Recalling cific points of difficulty and to design curricula with
Mathematics (Kinch) are examples of software de- these in mind. Errors are symptoms which allow
signed to help students prepare for the Entry Level us not only to identify stumbling blocks, but also
Mathematics Exam at California State University to form an impression of the students conceptions.
which have been favourably received. The computer allows students to respond to their
errors in. a new way: They can identify and con-
More and more, educational software includes a
trol them themselves. Getting rid of them can even
counsellor, helping the student to make his or her
become a motivation for learning.
way through the activities of the software, evaluat-
ing him or her, and individualising the activities. One example of the use of the computer to detect
and correct errors is found in Okon-Rinnes course-
(g) Assessment and Recording ware. This enables a student to choose a basic func-
tion such as f(z) = 1x1 and then to experiment with
The computer can be used for testing students the effects which translations and reflections have on
progress. Some software employs the random gen- it. Thus the graph can be translated vertically or
eration of test items. Such testing can, of course, horizontally or reflected in the vertical axis. Simul-
go far beyond reliance on multiple choice items and taneously the function changes to correspond to the
can measure responses other than correct and in- new graph. The intention is to detect such common
correct. Such newer testing procedures, which can errors as confusing J(z) = 12-21 with f(z) = l~+21,
be designed to capitalise on the graphic potentiali- or f(z) = I~+21 with f(z) = ltl+2. When an error
ties of the computer, can reduce testing time, allow is detected, a tutorial subroutine is activated and af-
tests to be broken off and resumed at any time, of- terwards the student has the option of continuing or
fer immediate summaries and analyses, and assign branching back to an earlier unit.
specific help for identified deficiencies.
The obvious disadvantages include preparation 3.4 Student responses to work with comput-
costs and the need to provide ready access to a com- ers
puter. Open-ended testing of projects or personal It is common to talk about the enthusiasm gen-
problem solving is at present difficult, but begin- erated in students by computer-baaed systems. In
nings are being made. Computer-assisted recording many experiments, it is claimed that this has re-
also has great potential. sulted in students developing a new interest in the
A computer at home, or a computer easily usable subject and that the general level of student activity
at school, enables students to use individualised sets had increased as a result of reacting with a computer
of data for homework or assessments package. Not only had activity increased, but so had
Very interesting examples of learning credit confidence. Dubinsky typically reports (of a course
cards are being experimented with: The card con- on discrete structures): This approach makes for
tains information about the learner, and gives him a lively course in which students are responsive in
or her access to appropriate software and activities. class and active outside class. In comparison with
similar groups to whom I have tried to teach this mathematicians are involved in creating such sys-
material, these students seem to be more prone to tems. As a result, sophisticated packages are self-
speak in terms of sets and less confused by compli- perpetuating, Since they will exist, we need to un-
cated logical statements. derstand their pedagogical uses and the possibly
It must not be thought, however, that enthusi- dramatic effects they could have on current mathe-
asm can be automatically generated through the use matics education.
of a computer. Much will depend on the students (b) Less sophisticated in computer terms but
and the teaching situation; there are also negative still very demanding in pedagogical design are the
experiences to report! One must also judge on how software packages suitable for use on a microcom-
much students learn as well as the enthusiasm they puter. These packages attempt to aid the stu-
show whilst engaged on the task. dents mathematical development and employ such
Here one is faced with a new problem in teach- themes as visualisation, simulation, exploration and
ing. Students can frequently appear fascinated by problem-solving. They may be used by students
computer demonstrations or by working interac- working alone, in groups, or with a teacher. Many
tively with a computer, but what happens when individuals and groups are writing such packages.
the machine is switched off? Will the students only Many are also provided by educational software
be able to imitate what they have seen or will they companies.
obtain a deeper understanding of concepts? A major problem arises here. The production of
It is recognised that the value of much computer packages that can be recommended for widespread
work is largely dependent upon the follow-up activ- use as pedagogically sound and well-tested is an
ities which must guard against the possibility that expensive, complicated task requiring considerable
the machine is doing all the work and providing all professional resources. It should involve fundamen-
the answers. Many traditional activities will still tal research based on the structured observation of
have to be carried out, thus suggesting yet again the materials in use in parallel with the develop-
that the computers main contribution will be to en- ment of the materials. Thus the team may need to
hance student understanding and not to save time include mathematicians, educators, psychologists,
for the lecturer. The introduction of the computer computer scientists, graphic designers, publishers
is unlikely to solve (or even ease) the problem of and editors. The financial needs of such a group
overloaded syllabuses. would be considerable.
(c) General purpose programming languages can
be used as tools aiding students mathematical de-
3.5 The provision of software velopment and are a readily available teaching re-
source. Extension of such languages or even cre-
The way software is conceived and designed ation of new ones expressly for this purpose would
evolves very quickly. The progress of technology be welcome.
and the development of multimedia tools enriches This brief discussion of the present position
the possibilities for pedagogical uses. The roles of points out the need (i) to establish channels of com-
teachers, of pedagogs and of computer scientists in munication so that researchers and educators are
software design has evolved. Very user-friendly sys- aware of resources currently available and (ii) to es-
tems allow any teacher to create teaching situations tablish structured research studies using currently
with computers. available resources in order to gain and share un-
Current software resources may be considered in derstanding of their use as pedagogical tools.
several categories: The emergence of software packages has raised
(a) Sophisticated systems (in computer terms) a new problem for mathematics teachers, that of
such as the symbolic manipulation systems, large black boxes, for they often/usually produce answers
statistical packages, etc., form the first category. without giving any hint of the way in which they
These systems have been developed in a goal- were obtained. This may well conceal a wealth of
oriented fashion, that is they seek to provide solu- deep mathematics. (It could, of course, be argued
tions to specific mathematical problems. They have that the problem is not new, but merely heightened -
not needed to consider to any great extent peda- for students have been employing algorithms whose
gogical design. Interest in their use as pedagogical workings they did not understand for centuries!).
tools is growing. How can students learn (be taught/encouraged)
Commercial companies exist with an interest to look critically at the answers supplied? How
in marketing this type of software and research much should students be required to know about
the workings of black-boxes before being allowed to Not all problems associated with computers in
use them? For example, there are packages which education can be anticipated. Many questions
invert matrices. If such a package uses floating-point need to be answered through research initiatives di-
arithmetic, it can give answers which should not be rected at investigating the possibilities, limitations
accepted at face value. At least students should be and possible dangers of computer use in education.
warned about this or, better, should learn to recog- Some causes for concern are:
nise when this has occured. uniformity in students thinking and reasoning
could arise from overuse of computers in the
3.6 Cultural, social and economic factors learning process,
standardisation of software development (in an
We have written of the computer as an aid to attempt to form a commercial market) may lead
mathematics teaching and learning. So is the over- to mediocrity and conformity,
head projector. The difference though between the subtleties of communication between teachers
two tools is not, however, solely the enormous dif- and students could be impoverished by over-
ference in the range of possibilities opened up by using computers,
the former. Equally, it springs from the enormous
insensitive working with computers could ad-
effect which the computer is having upon society
versely influence the total intellectual develop-
outside the confines of educational systems. As a
ment of students (of their intuitive thinking, cre-
result society has expectations concerning comput-
ativity, perception, etc.).
ers and their use, expectations which often have lit-
The case of developing countries demands spe-
tle basis in reality. Students too have expectations
cial attention. For them the provision and mainte-
about their use. There are then enormous pressures
nance of hardware creates great problems. More-
on educators at all levels to use computers, not nec-
over, scarce resources must be husbanded carefully.
essarily for their intrinsic value, but because society
The computer could offer special advantages to
expects it, and not because to do so might be con-
them; on the other hand the absence or shortage of
sidered old-fashioned and reactionary.
computers could widen still further the gap between
It will be difficult for computers to be used effec-
them and the developed countries. Several confer-
tively in education until society has become better
ences have considered the question of new technolo-
informed about their power and limitations. Un-
gies in education for developing countries (see, for
realistic expectations must be strongly discouraged.
example, Amara, Boudriga and Harzallah, 1986).
There is a danger that false advertising by computer
companies and software developers, and a pressure
from various sections of society, could lead to ill- 3.7 Conclusion
designed, over-optimistic innovation and, in turn,
We are only experiencing the very beginning of
to a backlash comparable with that of the 1970s
the effect of computers on the teaching and learn-
resulting from the hasty introduction of the New
ing of mathematics. Gradually, we are beginning
Math. Political decision makers in some countries
to take advantage of some of their more obvious
are pushing computers and computer-related cur-
possibilities such as their quick and accurate prc+
ricula into education without adequate considera-
duction of graphical material, their quick and accu-
tion of objectives and consequences.
rate (though not always precise) arithmetic, analyse
It is important, therefore, to realise that: large quantities of data.
l reasonable use of computers in education re- In numerous publications one can see examples
quires software programs and packages whose of mathematical situations for which the computer
educational standards and qualities are compa- and informatics allow us to see and approach situa-
rable with the technical ones offered by the avail- tions from a new point of view. Obvious examples
able hardware, which spring to mind are the many applications in
l integrating computers into the curriculum must statistics (dealing with vast quantities of data), in
be coordinated with teacher/faculty in-service, probability (with all the possibilities opened up for
professional development programs, simulation by pseuderandom number generators);
l educational budgets must be prepared to permit in geometry, too, there is a range of interesting ac-
appropriate expenditure on hardware, software, tivities - production and processing of images, curve
and teacher development, plotting, the transformation of images (translations,
l no curriculum should remain stagnant for a long reflections, etc.), loci, exploration of images and fig-
period. ures. The dynamic aspect dominates here: One can
visualise instantly the effect of varying a parameter. enthusiastic teachers have produced activities and
In linear algebra, an algorithmic approach furnishes tools, and have tried new pedagogical strategies us-
a tool both for calculations and also for demon- ing computers. We now need to have ALL teachers
stration. Here again the dynamic aspect plays an able to use computers as a natural tool, and to in-
important role: To see a matrix steadily assume a tegrate them into their teaching.
diagonal form is very different from obtaining the At the same time there is the need to carry out
result once and for all after a long and involved cal- much research and experimentation so that we may
culation. But it .is above all in analysis that the effectively understand and control the impact of the
opportunities to utilise informatics are richest and use of the computer on students learning and on
most numerous. The study of numbers, of func- their conceptions and representations of mathemat-
tions, of the solution of equations, observation and ical objects, Only after such studies will we be able
study of sequences and series (and in particular of to provide high quality software and, most impor-
their speed of convergence), integral calculus, differ- tantly, a new range of didactical activities, tasks and
ential equations, asymptotic expansions, discretisa- situations to enhance learning.
tion, power series for functions, etc. In addition to
these classical fields where the use of the computer REFERENCES
arises naturally, one has also seen developments in
newer fields which have occurred largely because of Banchoff, T. et al (Eds.) [1988]: Educational Com-
computers: Formal symbolic logic is a striking in- puting in Mathematics, ECM87, Amsterdam:
stance; discrete mathematics can provide us with North Holland.
other examples. The computer is not only an aid Amara, M., Boudriga, N. and Harzallah, K. (Eds.)
for computation and demonstration, but a force for [1986]: LInformatique et lenseignement des
development. mathdmatiques dans les pays en voie de
In all of these cases, the contribution of the com- ddveloppement, Proceedings of the first ICO-
puter takes several forms. Firstly, it is a calculating MIDC Symposium, Paris: ICOMIDC and UN-
tool allowing numerous and rapid calculations; it ESCO.
also serves to place renewed emphasis on numeri- Cornu, B. [1992]: LOrdinateur pour enseigner les
cal methods, and thus on the study of algorithms; mathematiques, Paris: Presses Universitaires
and, especially, it is a pedagogical tool for promot- de France.
ing teaching and learning.
Hirst, A. and Hirst, K. (Eds.) [1988]: Proceedings
However, let us reiterate, the act of using a com-
of the Sixth International Congress on Mathe-
puter does not automatically lead to an improve-
matics Education, Budapest, 1988, Budapest:
ment. It is not a magic wand! Like all tools, it can
Janos Bolyai Mathematical Society.
serve us badly; we must learn how to get the best
from it. Johnson, D.C. and Lovis, F. (Eds.) [1987]: Infor-
matics and the Teaching of Mathematics, Pro-
The development of technology (computers be-
ceedings of the IFIP TC 3/WG 3.1 Working
coming smaller and cheaper) and the development
Conference on Informatics and the Teaching
of new tools (such as multimedia ones) will certainly
of Mathematics, Sofia, Bulgaria, 16 - 18 May,
provoke very large changes in education. Complex
1987. Amsterdam: North-Holland. ISBN O-
learning environments and integrated software will
444-70325-X.
become more and more available. The technology
age in education is still to come! Tinsley, J.D. and van Weert, T.J. (Eds.) [1989]:
Computers are now widely to be found in schools Educational Software at the Secondary Level,
and universities, but they are not always widely Amsterdam: Elsevier.
used. Teachers are being trained in their use, but
principally in techniques and programming, and the
question of giving them a true pedagogical training
is not totally solved. It is also necessary to bear
in mind that if we wish to change the educational
system, then there will be a need simultaneously to
reform both the training given to those preparing to
teach in schools and universities and also the con-
tinuing education of existing teachers. Many inter-
esting and rich experiments have been done, many
LIVING WITH A NEW MATHEMATICAL SPECIES
Lynn Arthur Steen

St. Olaf College, Northfield, Minnesota 55077, U.S.A.
Computers are both the creature and the creator of mathematics has always been fallible; now it has
of mathematics. They are, in the apt phrase of Sey- a partner in fallibility.
mour Papert, mathematics-speaking beings. J. Research on the so-called Feigenbaum constant
David Bolter, in his stimulating book Turings Man reveals just how far this evolution has progressed in
[Bolter, 19841, calls computers embodied mathe- just a few years: computer-assisted investigations
matics. Computers shape and enhance the power of families of periodic maps suggested the presence
of mathematics, while mathematics shapes and en- of a mysterious universal limit, apparently indepen-
hances the power of computers. Each forces the dent of the particular family of maps. Subsequent
other to grow and change, creating, in Thomas theoretical investigations led to proofs that are true
Kuhns language, a new mathematical paradigm. hybrids of classical analysis and computer program-
Until recently, mathematics was a strictly hu- ming [Eckmann, 19841, showing that computer-
man endeavor. But suddenly, in a brief instant on assisted proofs are possible not just in graph theory,
the time scale of mathematics, a new species has but also in functional analysis.
entered the mathematical ecosystem. Computers Computers are also computing machines. By
speak mathematics, but in a dialect that is difficult absorbing, transforming, and summarizing massive
for some humans to understand. Their number sys- quantities of data, computers can simulate reality.
tems are finite rather than infinite; their addition is No longer need the scientist build an elaborate wind
not commutative; and they dont really understand tunnel or a scale model refinery in order to test en-
zero, not to speak of infinity. Nonetheless, they gineering designs. Wherever basic science is well
do embody mathematics. understood, computer models can emulate physical
The core of mathematics is changing under the processes by carrying out instead the process im-
ecological onslaught of mathematics-speaking com- plied by mathematical equations. Whereas mathe-
puters. New specialties in computational complex- matical models used to be primarily tools used by
ity, theory of algorithms, graph theory, and formal theoretical scientists to formulate general theories,
logic attest to the impact that computing is hav- now they are practical tools of enormous value in
ing on mathematical research. As Arthur Jaffe has the everyday world of engineering and economics.
argued so well (in [Jaffe, 1984]), the computer rev- It has been just over fifty years since Alan Turing
olution is a mathematical revolution. developed his seminal scheme of computability [Tur-
New Mathematics for a New Age ing, 19361 in which he argued that machines could
do whatever humans might hope to do. In abstract
Computers are discrete, finite machines. Unlike
terms, what he proposed was a universal machine of
a Turing machine with an infinite tape, real ma-
mathematics (see [Hodges, 19831 for details). It took
chines have limits of both time and space. Theirs is
two decades of engineering effort to turn Turings
not an idealistic Platonic mathemat,ics, but a math-
abstract ideas into productive real machines. Dur-
ematics of limited resources. The goal is not just to
ing that same period abstract mathematics flour-
get a result, but to get the best result for the least ef-
ished, led by Bourbaki, symbolized by the gener-
fort. Optimization, efficiency, speed, productivity-
alized abstract nonsense of category theory. But
these are essential objectives of modern computer
with abstraction came power, with rigor came cer-
mathematics.
tainty. Once real computers emerged, the complex-
Computers are also logic machines. They em-
ity of programs quickly overwhelmed the informal
body the fundamental engine of mathematics-
techniques of backyard programmers. Formal meth-
rigorous propositional calculus. The first celebrated
ods became de rigueur; even the once-maligned cat-
computer proof was that of the four-color theorem:
egory theory is now enlisted to represent finite au-
the computer served there as a sophisticated ac-
tomata and recursive functions (see, e.g., [Beckman,
countant, checking out thousands of cases of reduc-
19841, [Lewis, 19811). 0 nce again, as happened be-
tions. Despite philosophical alarms that computer-
fore with physics, mathematics became more effica-
based proofs change mathematics from an a pri-
cious by becoming more abstract.
ori to a contingent, fallible subject (see, e.g., [Ty-
moczko, 1979]), careful analysis reveals that noth-
ing much has really changed. The human practice
33
The Core of the Curriculum were created and over 40 new textbooks in discrete
Twenty-five years ago in the United States mathematics were published.
the Committee on the Undergraduate Program in Soon calculus itself came under scrutiny, as a
Mathematics (CUPM) issued a series of reports that natural force for counter-reformation. Critics ar-
led to a gradual standardization of curricula among gued that the power of computation and the ubiq-
undergraduate mathematics departments [CUPM, uity of applications had changed fundamentally the
19651. Shortly thereafter, in 1971, Garrett Birkhoff role of calculus in the practice of mathematics (e.g.,
and J. Barkley Rosser presented papers at a meeting [Douglas, 1986; Steen, 19881). The National Science
of the Mathematical Association of America con- Foundation launched diverse projects to reshape the
cerning predictions for undergraduate mathemat- nature of calculus instruction. Virtually all of these
ics in 1984. Birkhoff urged increased emphasis on projects feature supporting roles for the numeric,
modelling, numerical algebra, scientific computing, symbolic, and graphic power of computers.
and discrete mathematics. He also advocated in- The need for consensus on the contents of un-
creased use of computer methods in pure math- dergraduate mathematics is perhaps the most im-
ematics: Far from muddying the limpid waters portant issue facing American college and univer-
of clear mathematical thinking, [computers] make sity mathematics departments [CUPM, 19891. On
them more transparent by filtering out most of the the one hand departments that have a strong tra-
messy drudgery which would otherwise accompany ditional major often fail to provide their students
the working out of specific illustrations. [Birkhoff, with the robust background required to survive the
1972, p. 6511 R osser emphasized many of the same evolutionary turmoil in the mathematical sciences.
points, and warned of impending disaster to un- Like the Giant Panda, these departments depend for
dergraduate mathematics if their advice went un- survival on a dwindling supply of bamboo-strong
heeded: Unless we revise [mathematics courses] students interested in pure mathematics. On the
so as to embody much use of computers, most of other hand, departments offering flabby composite
the clientele for these courses will instead be taking majors run a different risk: by avoiding advanced,
computer courses in 1984. [Rosser, 1972, p. 6391 abstract requirements, they often misrepresent the
In the first decade after these words were writ- true source of mathematical knowledge and power.
ten, U.S. undergraduate and graduate degrees in Like zoo-bred animals unable to forage in the wild,
mathematics declined by 50%. The clientele for students who have never been required to master
traditional mathematics migrated to computer sci- a deep theorem are ill-equipped to master the sig-
ence, and the former CUPM consensus all but dis- nificant theoretical complications of real-world com-
appeared. In 1981 CUPM issued a new report, this puting and mathematics.
one on the Undergraduate Program in Mathemat-
Computer Literacy
ical Sciences ([CUPM, 19811, reprinted in [CUPM,
19891). Beyond calculus and linear algebra, they Mathematical scientists at American institutions
could agree on no specific content for the core of a of higher education are responsible not only for the
mathematics major: There is no longer a common technical training of future scientists and engineers,
body of pure mathematical information that every but also for the technological literacy of the edu-
[mathematics major] should know. cated public-of future lawyers, politicians, doctors,
The symbol of reformation became discrete educators, and clergy. Public demand that college
mathematics. Anthony Ralston argued forcefully graduates be prepared to live and work in a com-
the need for change before both the mathematics puter age has caused many institutions to introduce
community [Ralston, 19811 and the computer sci- requirements in quantitative or computer literacy.
ence community [Ralston, 19801. Discret.e math- In 1981 the Alfred P. Sloan Foundation initiated
ematics, in Ralstons view, is the central link be- curricular exploration of the new liberal arts, the
tween the fields. The advocacy of discrete math- role of applied mathematical and computer sciences
ematics rapidly became quite vigorous (see, e.g., in the education of students outside technical fields.
[Kemeny, 19831, [Ralston, 1983,] and [Steen, 1984]), The ability to cast ones thoughts in a form that
and the Sloan Foundation funded experimental cur- makes possible mathematical manipulation and to
ricula at six institutions to encourage development perform that manipulation . . . [has] become essen-
of discrete-based alternatives to st,andard freshman tial in higher education, and above all in liberal ed-
calculus. The impact of this work can be seen in ucation. [Koerner, 1981, p, 61 Ot,hers echoed this
the growth of courses and publicat,ions: in the five call for reform of liberal education. David Saxon,
year period from 1985 to 1990, hundreds of courses President of the University of California wrote in
A New Mathematical Species 35
a Science editorial that liberal education will con- cover a broad and varied spectrum, from business-
tinue to be a failed idea aa along as our students oriented data processing curricula, through manage-
are shut out from, or only superficially acquainted ment information science, to theoretical computer
with, knowledge of the kinds of questions science science. All of these intersect with the mathematics
can answer and those it cannot. [Saxon, 19821 curriculum, each in different, ways.
Too often these days the general public views To help clarify these conflicting approaches,
computer literacy as a modern substitute for math- Mary Shaw of Carnegie Mellon University put to-
ematical knowledge. Unfortunately, this often leads gether a composite report on the undergraduate
students to superficial courses that emphasize vo- computer science curriculum. This report is quite
cabulary and experiences over concepts and princi- forceful about the contribution mathematics makes
ples [Steen, 1985). The advocates of computer lit- to the study of computer science: The most im-
eracy conjure images of an electronic society dom- portant contribution a mathematics curriculum can
inated by the information industries. Their slogan make to computer science is the one least likely to
of literacy echoes traditional educational values, be encapsulated as an individual course: a deep ap-
conferring the aura but not the logic of legitimacy. preciation of the modes of thought that characterize
Typical courses in computer literacy are filled mathematics. [Shaw, 1984, p. 551
with ephemeral details whose intellectual life will The converse is equally true: one of the more
barely survive the students school years. These important contributions that computer science can
courses contain neither a Shakespeare nor a Newton, make to the study of mathematics is to develop in
neither a Faulkner nor a Darwin; they convey no students an appreciation for the power of abstract
fundamental principles nor enduring truths. Com- methods when applied to concrete situations. Stu-
puter literacy is more like driver education than like dents of traditional mathematics used to study a
calculus. It teaches students the prevailing rules of subject called Real and Abstract Analysis; stu-
the road concerning computers, but does not leave dents of computer science now can take a course
them well-prepared for a lifetime of work in the in- titled Real and Abstract Machines. In the for-
formation age. mer new math, as well as in modern algebra, stu-
Algorithms and data structures are to computer dents learned about relations, abstract versions of
science what functions and matrices are to math- functions; today business students study relational
ematics. As much of the traditional mathematics data structures in their computer courses, and ad-
curriculum is devoted to elementary functions and vertisers tout fully relational as the latest innova-
matrices, so beginning courses in computing--by tion in business software.
whatever name--should stress standard algorithms An interesting and pedagogically attractive ex-
and typical data structures. As early as students ample of the power of abstraction made concrete can
study linear equations they could also learn about be seen in the popular electronic spreadsheets that
stacks and queues; when they move on to conic sec- are marketed under such trade names as Lotus and
tions and quadratic equations, they could in a par- Excel. Originally designed for accounting, they can
allel course investigate linked lists and binary trees. as well emulate cellular automata or the Ising model
Computer languages can (and should) be stud- for ferromagnetic materials [Hayes, 19831. They
ied for the concepts they represent-procedures in can also be programmed to carry out most stan-
Pascal and C, recursion and lists in Lisp-rather dard mathematical algorithms--the Euclidean al-
than for the syntactic details of semicolons and line gorithm, the simplex method, Eulers method for
numbers. They should not be undersold as mere solving differential equations [Arganbright, 19851.
technical devices for encoding problems for a dumb An electronic spreadsheet--the archetype of ap-
machine, nor oversold as exemplars of a new form plied computing-is a structured form for recur-
of literacy. Computer languages are not modern sive procedures--the fundamental tool of algorith-
equivalents of Latin or French; they do not deal in mic mathematics. It is a realization of abstract
nuance and emotion, nor are they capable of per- mathematics, and carries with it much of the power
suasion, conviction, or humor. Although computer and versatility of mathematics.
languages do represent a new and powerful way to
think about problems, they are not a new form of Computers in the Classroom
literacy. Just as the introduction of calculators upset the
comfortable pattern of primary school arithmetic,
Computer Science
so the spread of computers will upset the traditions
In the United States, computer science programs of secondary and tertiary mathematics. This year
long division is passe; next year integration will be of rules and template-driven tests (see e.g., [Smith,
under attack. 1988; Zorn, 19871).
The impact of computing on secondary school It is commonplace now to debate the wisdom
mathematics has been the subject of many discus- of teaching skills (such as differentiation) that com-
sions in the United States (e.g., [Steen, 19871). Jim puters can do as well or better than humans. Is
Fey, coordinator of two assessments ([Corbitt, 1985; it really worth spending one month of every year
Fey, 1984]), described these efforts as an unequivc+ teaching half of a countrys 18-year-old students
cal dissent from the spirit and substance of efforts to how to imitate a computer? What is not yet so
improve school mathematics that seek broad agree- common is to examine critically the effect of ap-
ment on conservative curricula. [Fey, 1984, p. viii] plying to mathematics pedagogy computer systems
The new Curriculum and Evaluation Standards for that are themselves only capable of following rules
School Mathematics [NCTM, 19891 of the National or matching templates. Is it wise to devise sophisti-
Council of Teachers of Mathematics as well as other cated computer systems to teach efficiently precisely
recommendations from the U.S. National Academy those skills that computers can do better than hu-
of Sciences ([NRC, 1989; MSEB, 19901) set expec- mans, particularly those skills that make the com-
tations for school mathematics that employ calcula- puter tutor possible? In other words, since com-
tors and computers in every appropriate manner. puters can now do the calculations of algebra and
Teachers in tune with the computer age seek calculus, should we use this power to reduce the
change in both curriculum and pedagogy. But the curricular emphasis on calculations or to make the
inertia of the system remains high. For example, teaching of these calculations more efficient? This
the 1982 International Assessment of Mathematics is a new question, with a very old answer.
documented that in the United States calculators
Let Us Teach Guessing
are never permitted in one-third of the 8th grade
classes, and rarely used in all but 5% of the rest Forty years ago George P6lya wrote a brief pa-
[McKnight, 19871. Recent data [NAEP, 19911 show per with the memorable title Let Us Teach Guess-
some improvement, but still fall well short of the ing [Pblya, 19501. It is not differentiation that our
NCTM recommendations. students need to learn, but the art of guessing. A
Laptop computers are now common-they cost month spent learning the rules of differentiation re-
about as much as ten textbooks, but take up only inforces a students ability to learn (and live by)
the space of one. Herb Wilf argues (in [Wilf, 19821) the rules. In contrast, time spent making conjec-
that it is only a matter of time before students will tures about derivatives will teach a student some-
carry with them a device to perform all the al- thing about the art of mathematics and the science
gorithms of undergraduate mathematics. Richard of order.
Rand, in a survey [Rand, 19841 of applied research With the aid of the mathematics-speaking com-
based on symbolic algebra agrees: It will not be puter, students can for the first time learn college
long before computer algebra is as common to engi- mathematics by discovery. This is an opportunity
neering students as the now obsolete slide rule once for pedagogy that mathematics educators cannot af-
was. Just five years after Wilfs article appeared, ford to pass up. Mathematics is, after all, the sci-
the same journal carried a review [Nievergelt, 19871 ence of order and pattern, not just a mechanism for
of the first pocket calculator with symbolic algebra grinding out formulas. Students discovering math-
capabilities. ematics gain insight into the discovery of pattern,
Widespread use of computers that do school and and slowly build confidence in their own ability to
college mathematics will challenge standard educa- understand mathematics. Formerly, only students
tional practice [Steen, 19901. For the most part, of sufficient genius to forge ahead on their own could
computers reinforce the students desire for cor- have the experience of discovery. Now with comput-
rect answers. In the past, their school uses have ers as an aid, the majority of students can experi-
primarily extended the older teaching machines: ence for themselves the joy of discovery.
programmed drill with pre-determined branches for
Metaphors for Mathematics
all possible responses. But the recent linking
of symbolic algebra programs with so-called ex- Two metaphors from science are useful for un-
pert systems into sophisticated intelligent t,utors derstanding the relation between computer science,
has produced a rich new territory for imaginative mat,hematics, and education. Cosmologists long de-
computer-assisted pedagogy that advocates claim bat,ed two theories for the origin of the universe-
can rescue mathematics teaching from the morass the Big Bang theory, and the theory of Continuous
A New Mathematical Species 37
Creation. Current evidence tilts the cosmology de- tions of Electronic Spreadsheets, New York:
bate in favor of the Big Bang. Unfortunately, this McGraw-Hill.
is all too often the public image of mathematics as
Beckman, F.S. [1984]: Mathematical Foundations of
well, even though in mathematics the evidence fa-
Programming. The Systems Programming Se-
vors Continuous Creation.
ries, Reading, MA: Addison-Wesley.
The impact of computer science on mathemat-
ics and of mathematics on computer science is the Birkhoff, G. [1972]: The Impact of Computers
most powerful evidence available to beginning stu- on Undergraduate Mathematics Education in
dents that mathematics is not just the product of 1984, Amer. Math. Monthly, 79, 648-657.
an original Euclidean big bang, but is continually
Bolter, J.D. [1984]: Turings Man: Western Cul-
created in response to challenges both internal and
ture in the Computer Age, Chapel Hill, NC:
external. Students today, even beginning students,
University of North Carolina Press.
can learn things that were simply not known twenty
years ago. We must not only teach new mathemat- Committee on the Undergraduate Program in
ics and new computer science, but we must teach as Mathematics [1965]: A General Curriculum
well the fact that this mathematics and computer in Mathematics for Colleges, Washington, DC:
science is new. Thats a very important lesson for Mathematical Association of America.
the public to learn.
Committee on the Undergraduate Program in
The other apt metaphor for mathematics comes
Mathematics [1981]: Recommendations for
from the history of the theory of evolution. Prior
a General Mathematical Sciences Program,
to Darwin, the educated public believed that forms
Washington, DC: Mathematical Association of
of life were static, just a~ the educated public of
America.
today assumes that the forms of mathematics are
static, laid down by Euclid, Newton, and Einstein. Committee on the Undergraduate Program in
Students learning mathematics from contemporary Mathematics [1989]: Reshaping College Math-
textbooks are like the pupils of Linnaeus, the great ematics. MAA Notes No. 13, Washington,
eighteenth-century Swedish botanist: they see a DC: Mathematical Association of America.
static, pre-Darwinian discipline that is neither grow-
Corbitt, M.K., and Fey, J.T. (Eds.) [1985]: The
ing nor evolving. Learning mathematics for most
Impact of Computing Technology on School
students is an exercise in classification and memo-
Mathematics: Report of an NCTM Confer-
rization, in labeling notations, definitions, theorems,
ence, Reston, VA: National Council of Teach-
and techniques that are laid out in textbooks as so
ers of Mathematics.
much flora in a wondrous if somewhat abstract Pla-
tonic universe. Douglas, R.G. (Ed.) [1986]: Toward A Lean
Students rarely realize that mathematics con- and Lively Calculus. MAA Notes Number 6,
tinually evolves in response to both internal and Washington, DC: Mathematical Association of
external pressures. Notations change; conjectures America.
emerge; theorems are proved; counterexamples are
Eckmann, J.-P., Koch, H. and Wittwer, P. [1984]:
discovered. Indeed, the passion for intellectual or-
A Computer-assisted Proof of Universality for
der combined with the pressure of new problems- Area-preserving Maps, Memoirs of the Amer-
especially those posed by the computer-force re-
ican Mathematical Society, 47:289 (January).
searchers to continually create new mathematics
and archive old theories. Fey, J.T., et al. (Eds.) [1984]: Computing
The practice of computing and the theory of and Mathematics: The Impact on Secondary
computer science combine to change mathematics School Curricula, Reston, VA: National Coun-
in ways that are highly visible and attractive to stu- cil of Teachers of Mathematics.
dents. This continual change reveals to students the Hayes, B. [1983]: Computer Recreations, Scientific
living character of mathematics, restoring to the ed- American, 22-36, (October).
ucated public some of what the experts have always
known-that mathematics is a living, evolving com- Hodges, A. [1983]: Alan Turing: The Enigma, New
ponent of human culture. York: Simon and Schuster.
REFERENCES Jaffe, A. [1984]: 0 r d ering the Universe: The Role of

Mathematics in Renewing U.S. Mathematics,
Arganbright, D.E. [1985]: Mathematical Applica- Washington, DC: National Academy Press.
Kemeny, J.G. [1983]: Finite Mathematics-Then Rand, R.H. [1984]: Computer Algebra in Ap-
and Now in Ralston, A., and Young, G.S. plied Mathematics: An Introduction to MAC-
(Eds.), The Future of College Mathematics, SYMA, London: Pitman.
pp. 201-208, New York: Springer-Verlag. Rosser, J.B. [1972]: Mathematics Courses in 1984,
Koerner, J.D. (Ed.) [1981]: The New Liberal Arts: Amer. Math. Monthly, 79, 635-648.
An Exchange of Views, New York: Alfred P. Saxon, D.S. [1982]: Liberal Education in a Techno-
Sloan Foundation. logical Age, Science, 218, 845 (26 November).
Lewis, H.R., and Papadimitriou, C.H. [1981]: Ele- Shaw, M. (Ed.) [1984]: The Carnegie-Mellon Cur-
ments of the Theory of Computation, Engle- riculum for Undergraduate Computer Science,
wood Cliffs, NJ: Prentice-Hall. New York: Springer-Verlag.
Mathematical Sciences Education Board [1990]: Re- Smith, D.A.; Porter, G.J.; Leinbach, L.C.; and
shaping School Mathematics: A Philosophy Wenger, Ronald H. (Eds.) [1988]: Computers
and Framework for Curriculum, Washington, and Mathematics: The Use of Computers in
DC: National Academy Press. Undergraduate Instruction. MAA Notes No.
9, Washington, DC: Mathematical Association
McKnight, C.C., et al [1987]: The Underachiev- of America.
ing Curriculum: Assessing U.S. School Math-
Steen, L.A. [1984]: 1 + 1 = 0: New Math for a New
ematics from an International Perspective,
Age, Science, 225, 981 (7 September).
Champaign, IL: Stipes.
Steen, L.A. [1991]: T wenty Questions for Computer
National Assessment of Education Progress [1991]: Reformers in Demana, F., et. al. Proceedings
The State of Mathematics Achievement, of the Conference on Technology in Collegiate
Washington, DC: U.S. Department of Educa- Mathematics, Reading, MA: Addison-Wesley,
tion. pp. 16-19.
National Council of Teachers of Mathematics [1989]: Steen, L.A. (Ed.) [1988]: Calculus for a New Cen-
Cumculum and Evaluation Standards for tury: A Pump, Not a Filter, Washington, DC:
School Mathematics, Reston, VA: National Mathematical Association of America.
Council of Teachers of Mathematics. Steen, L.A. [1987]: Who Still Does Math with Paper
National Research Council [1989]: Everybody and Pencil?, Chronicle of Higher Education,
Counts: A Report to the Nation on the Fu- A48 (14 October).
ture of Mathematics Education, Washington, Steen, L.A. [1985]: Mathematics and Computing
DC: National Academy Press. Education: A Common Cause, Communica-
tions of the ACM, 28, 666-667.
Nievergelt, Y. [1987]: The Chip with the Col-
lege Education: The HP-28C, Amer. Math. Turing, A.M. [1936]: On Computable Numbers,
Monthly, 94, 895-902. with an Application to the Entscheidungs-
problem, Proc. London Math. Sot., 2nd Ser.,
Polya, G. [1950]: Let Us Teach Guessing, Etudes de 42, 230-265.
Philosophie des Sciences, Neuchatel: Griffon,
pp. 147-154; reprinted in G. Polya, Collected Tymoczko, T. [1979]: The Four Color Problem and
Papers, Vol. IV, Cambridge, MA: MIT Press, its Philosophical Significance, Journal of Phi-
1984, pp. 504-511. losophy, 76, 57-85.
Wilf, H.S. [1982]: The Disk with the College Edu-
Ralston, A. [1981]: Computer Science, Mathemat-
cation, Amer. Math. Monthly, 89, 4-8.
ics, and the Undergraduate Curricula in Both,
Amer. Math. Monthly, 88, 472-485. Zorn, P. [1987]: Computing in Undergraduate
Mathematics, Notices of the American Math-
Ralston, A., and Shaw, M. [1980]: Curriculum 78: ematical Society, 34, 917-923 (October).
Is Computer Science Really that Unmathe-
matical?, Communications of the ACM, 23,
67-70.
Ralston, A., and Young, G.S. (Eds.) [1983]: The

Future of College Mathematics, New York:
Springer-Verlag.
WHAT ARE ALGORITHMS? WHAT IS ALGORITHMICS?
Stephen B. Maurer
Swarthmore College, Swarthmore, Pennsylvania 19081-1397, U.S.A.
Overview. Roughly speaking, an algorithm is a problem for the given input, or it must indi-
precise, systematic method for solving some class of cate that it cannot solve the problem.
problems. Algorithmics is the systematic study of In some cases it is reasonable to relax these stringent
algorithms - how to devise them, describe them, val- requirements; well take up this point later. One
idate them and compare their relative merits. There can also ask: how precise is precise? Just how are
have been algorithms in mathematics since ancient the rules to be stated to make them precise? Good
times, but algorithmics is new. Only with the ad- question. It depends on who or what you are talking
vent of computers has it been possible to tackle such to. We will also address this further. But lets turn
large and complicated problems that a systematic immediately to some examples.
approach to algorithms is necessary. Because al-
gorithms are now essential in almost all business Example 1: Arabic Multiplication
and scientific applications of mathematics (as well The traditional paper and pencil algorithm for
as being increasingly important to mathematicians multiplying two numbers expressed in arabic numer-
themselves and fundamentally important to com- als is brilliant. Too bad we all take it for granted.
puter scientists), it is important that mathematics Its brilliant because it reduces a general problem
education take algorithms and algorithmics into ac- to a small subcase - how to multiply two single-
count. digit integers - and does so in a small amount of
This paper has four sections. In Section 1, by far space. Heres the result of applying the algorithm
the longest, we explain what algorithms are in much to 432 x 378:
more detail, presenting many examples. In Section
378
2 we do the same for algorithmics. In Section 3 we
discuss several reasons why the study of algorithms 3456
and algorithmics is valuable in mathematics, and 3024
we also discuss some counterarguments. Finally, in 1296
Section 4 we make some suggestions for incorporat-
ing algorithms and algorithmics into the secondary 163296
and tertiary mathematics curriculum. Each row of intermediate calculation is obtained
by multiplying the top factor (432) by one digit of
1. What Are Algorithms?
the bottom factor. If we expand out the first inter-
Algorithms turn input data into output data mediate row in more detail, we get
through sequences of actions. For instance, an al-
432
gorithm might take two integers and output their
8
product. The rules specifying the algorithm (includ-
ing rules specifying what inputs are allowed) must 16
be precise enough to satisfy 24 (1)
1. Determinateness. For each allowed input, 32
the first action is uniquely determined, and
3456
more generally, after each action in the se-
quence the successor action is uniquely de- Of course, its never written this way. To save
termined. space, the carries are either all done mentally, or
It doesnt do us any good to have an algorithm that they are marked with small digits as follows:
doesnt stop, so we also require
2. Finiteness. For any allowed input, the al- 3 42516
gorithm must stop after a finite sequence of We include Display (1) t o make the role of single-
actions. digit multiplications explicit. For instance, 16 is the
Usually algorithms are devised to solve problems. product of the 2 in 432 by the 8 in 378.
Such algorithms must be appropriate for the pur- Now, is this format precise enough for present-
pose at hand: ing Arabic multiplication? Apparently so, because
3. Conclusiveness. When the algorithm termi- such a format does seem to suffice for teaching the
nates, it must either output a solution to the algorithm to children (when presented with many
39
examples, lots of oral explanations, and hands-on subtle. (For instance, we dont include a step just
practice). And you do need to make use of dia- before endfor to carry Al, because at this point Ai
grams if the physical positioning of symbols on the is always 0. Do you see why?)
page is part of the algorithm. The advantage of this formulation of the algo-
Nonetheless, this is not the format we will use rithm is that it is easier to verify that it is an al-
for other algorithms, and it is not a good format gorithm. Is it determinate? Yes, because each line
for systematically .verifying the defining conditions leaves no doubt about what is to be done, and the
l-3 above, So we now restate Arabic multiplica- order of execution is also specified - go down the
tion using algorithmic language, a language style page, except when you get to the end of a loop, go
quite similar to a programming language. (We will back to the beginning. Is it finite? Yes, because
assume basic familiarity with how such languages the loop has only 5 lines, and the loop gets carried
are to be read, e.g., what a loop is, what an assign- out m + 1 times. Does it solve the problem? This
ment is.) Actually, we restate only the part shown is not so obvious, but the specificity of the lines
in Display (1) - a multidigit number times a sin- makes it easier to present a proof when it is time to
gle digit number The algorithm makes use of two get around to that. (We will talk about algorithm
procedures, verification later.)
DigMult(a, b) which multiplies the single Notice that this algorithm involves iteration:
digits a and b and returns Ml some subprocess is applied repetitively. In this case
and M,., the left and right dig- the subprocess of multiplying two single-digit num-
its of the product. bers (and then carrying) was iterated. While an
algorithm does not have to involve iteration (or a
DigAdd(u, 6) which adds the single digits a
related type of repetition called recursion), almost
and b and returns Al and A,,
all algorithms of interest in mathematics do.
the left and right digits of the
sum. (Al will be either 0 or Example 2: Euclids Algorithm
1.)
This one is much older than the first, and also
Heres the algorithm: much simpler, but perhaps not so well known. It is
the classical Greek method for finding the greatest
Input ao, al,. , a,,, [the ones, tens,. . , digits common divisor (gcd) of two positive integers. It
of an (m+l)-digit number] assumes you already know how to divide and find
b [the one-digit multiplier] remainders. The algorithm keeps dividing and find-
Algorithm ing a remainder until the remainder is 0. Then the
carry +-- 0 previous remainder is the gcd of the original num-
for j = 0 to m bers.
DigMult(aj, 6) Here is a numerical example. Find the gcd of
DigAdd(M,, carry) [add any carry from 147 and 33. The quotient of 147 divided by 33 is 4
previous product] with remainder 15. That is,
Pj c A,. [jth digit of the product known] 147 = 33 x 4 + 15.
DigAdd(Ml, Al) [needed in case the carry
So any number that divides 147 and 33 also divides
affects the left digit]
15, and conversely, any number that divides 33 and
carry + A,. [carry to the next single-
15 divides 147. Now, do the same operations to 33
digit multiplication] and 15 that we did to 147 and 33: 15 divides into 33
endfor with remainder 3. Thus a number divides 33 and 15
if carry > 0 then Z,+l +- carry if and only if it divides 15 and 3. But 3 divides into
output PO, P,, . . , P,,, and sometimes Pm+1 15 exactly. So the largest number dividing 3 and 15
[digits of the product] is 3 itself. Thus the gcd of 147 and 33 is 3.
In algorithmic language, Euclids algorithm is
This is no doubt hard to follow, but try carry- the following:
ing it out on the example above. Look at 432 x 7
(the middle line of the first example), which shows
why the two lines before endfor are needed. That
this description is hard to follow should bring home
the point that the Arabic algorit,hm is really quite
Algorithms and Algorithmics 41
Input rn, n [integers >= 0]

Algorithm
num + m; denom + n
repeat until denom = 0 In particular, the (2,l) entry of the product is
quot +- [num/denomJ
4*7+5*(-5)+6*3=21
[integer part of num/denom]
rem + num - quot*denom How can we express the definition of matrix multi-
num + denom; denom + rem plication as an algorithm?
[update num and denom] Informally, you just go through each combina-
endrepeat tion of a row from A and a column from B and com-
Output num pute their product according to (2). Their product
is a sum of real-number products, so we can com-
pute it by keeping a running sum and successively
For instance, for the numerical example above,
adding real products until we are done. In algorith-
initially num(erator) is 147 and denom(inator) is 33.
mic language we have
Since 33 # 0, we enter the repeat loop, quot(ient)
is computed as 4 and rem(ainder) as 15. Then
(33,15) become the new (num,denom) pair. Since Input A, B, m, n,p
denom is still not 0, we traverse the loop again, and Algorithm
(num,denom) becomes (15,3). At this point, work- for i = 1 to m
ing by hand, we immediately recognized that 3 di- for L = 1 to p
vides 15, but a computer must discover this by Cik + 0
following the rules. Since 3 # 0, we enter the loop [initialize the ilc entry of C = Al?]
again, and update (num,denom) to (3,0). Now de- for j = 1 to 71
nom = 0 and the algorithm quits, outputting num = cik + Cik + Uij * bjk
3 as the gtd. endfor
Notice there is no factoring in this algorithm. endfor
Another way to find gcd(m, n) is to factor m and endfor
n, and then take the product of all common factors.
output c
This second method is the standard one currently
taught in elementary schools in North America. For
Example 4: Construct fi
small values of m and n, the second method is often
faster than Euclids method, but factoring very large All the examples so far have been arithmetic or
numbers is very hard. In general, Euclids method algebraic. Heres one from geometry. By construct-
is the way to go. ing a number r, we shall mean constructing a line
Euclids method is an algorithm. Clearly it is segment of length r, starting with a line segment of
determinate. It is finite, because rem is always a length 1 and using a straightedge and compass. To
nonnegative integer and gets smaller with each it- construct a, construct a unit perpendicular at one
eration, so eventually it must reach 0 and the algo- end of the initial unit segment. By the Pythagorean
rithm stops. The algorithm is conclusive (correctly Theorem, the hypotenuse has length 4. Now it is
determines the gcd) for the reasons we argued in- possible to construct fi by repeating the process.
formally above. A formal proof would be by math- Construct a unit perpendicular at the end of the seg-
ematical induction. ment of length fi. The new hypotenuse will have
length A. See Figure 1.
1
Example 3: Matrix Multiplication
1
Let A be an m x n matrix and B an n x p matrix.
Call the entry of A in row i (down from the top) and
column j (from the left) aij. Similarly, B = [bjk].
Then their product Al3 is defined to be the m x p A=2
xh 6
matrix whose (i, L) entry is
n
j=l
Lb 1
1
For instance, Figure 1

By induction, it should be clear that fi may be

constructed for all positive integers IZ. Here is the
construction in algorithmic language.
Input n, unit line segment AB

Algorithm
for c = 2 to n
Construct BC I AB, with BC = 1
AB+AC [change names]
endfor
Output AB [segment of length fl
Figure 2
Is this determinate enough to be called an algo-
rithm? It depends on the audience. If the reader
knows well how to construct perpendiculars with
Input num, Pinit, Pfin [number of disks,
straightedge and compass, it is. If not, the line
Construct BC I Al? must be expanded. initial pole number, final pole number]
Algorithm
Example 5: Towers of Hanoi procedure H(n, r, s)
[move n disks from pole r to pole s]
Algorithmic approaches apply not just to tradi- if n = 1 then Move disk on T to s
tional ma.thematical topics, but also to any situa- else H(n- 1, r, 6-r-s)
tion where a systematic and repetitive approach is [move all but bottom disk to nontarget pole]
needed for a solution. Towers of Hanoi (TOH) is a Move disk on r to s
game played with a set of n rings (or disks) of differ- H(n-1,6-r-s, s)
ent sizes and three poles. Initially the rings are all [move other disks onto target pole]
on one pole, from smallest on top to largest on the endif
bottom. The object is to get them all to another
endprocedure
pole, in the same order, making moves according to
H(num, Pinit, Pfin)
the following rules.
[main algorit,hm - invoke H]
1. Move only one ring at a time.
Output Solution to the game
2. A larger ring may never be placed on a
smaller ring. That this is an algorithm is not so clear. Its
TOH is often used by psychologists doing experi- not clear how to start carrying out the call of H,
ments wit,h children. While it is easy to figure out since mostly it just calls itself again instead of mov-
solutions for n = 3 or 4, for larger R most kids soon ing disks. Its also not clear that when it, finishes
lose their way. University students often dont do (if it finishes), it has solved the game. But in fact
much better! The key to understanding why the it is an algorithm, and once one develops a good
game can be solved is recurszon - reduce to the pre- understanding of how recursion works, it is fairly
vious case. Suppose we already know how to solve evident why. In any event, good programming lan-
the (n-l)-ring game. Regarding that subgame a.5 guages have recursion built in, and t,hus the algo-
an indivisible block, then Figure 2 shows how to rithm above is easy to translat,e into such la.nguages.
solve the n-ring game. This solution may be put into
Example 6: The Quadratic formula
algorithmic language if we allow a procedure (recall
DigMult in Example 1) to invoke itself. The pro- The traditional formula for solving az+bz+c =
cedure H in the algorithm is first defined (in terms 0 seems simple enough; wheres the algorithm and
of itself) and t.hen invoked by the (one-line) main why bother with it? Well, there are several cases -
algorithm. The poles are numbered 1,2,3. Note, two distinct real roots, one repeat,cd real root, no
therefore, that, if r and s are numbers of two differ- real roots ~ and properly choosing between cases is
ent, poles, then 6 - r - s is the number of the third an algorithmic matter. Even if the audience knows
pole. about complex numbers, if t,hey want, to c0mput.e
solut,ions, there is the problem t,hat, most calculators Input a, b [f(a)f(b) < 01
and comput,ers wont accept a request to take the Algorithm
square root of a negative number. So presenting the repeat
solution process as an algorithm has merit. c c- (a + b)/2
if J(c) = 0 then exit
Input a, b, c [coefficients of if sign(f(c)) = sign(f(a))
oc2 + bz + c, with a # 0] then b + c
Algorithm else a + c
D + b2 - 4ac endrepeat
if [three cases follow] output c
D > 0 then [two real roots]
S+-fi Now, this is not an algorithm, because it can go
ZI - (b+s)/2a on forever. For instance, if f(z) = z2 - 2, a = 1 and
x2 e (b-s)/2a b = 2, then it takes an infinite number of halvings
D = 0 then 11 -22 + b/2a to converge to the root c = fi. Of course, a cutoff
[one repeated real root] condition can be added:
D < 0 then [two complex roots] endrepeat when la-b1 < tolerance
St-&D
for whatever tolerance you choose. Even with such
~1 +- (b+is)/2a
a condition, a real computer running this algorithm
x2 + (b-is)/2a may not terminate, because, if the tolerance chosen
endif is very small, roundoff error may result in la-61 >
Output the roots, 21 and zz tolerance no matter how many iterations are per-
formed.
If we want to be even more comprehensive, and Nonetheless, it may be best to present this algo-
allow input with a = 0, then we have to include rithm initially in the nonterminating form above -
several more cases. Note that there are no loops in it gets at the key idea of bisection without obscur-
this algorithm, but several if-statements (even more ing details, and it also ties in with the concepts of
if a = 0 is allowed). Many procedures in the ev- infinite processes and limits needed for a full math-
eryday world involve more multiple decisions than ematical attack. So this is our first example that
iteration - think of tax laws. Such procedures trans- suggests why the three defining conditions at the
late into algorithms with many if-statements. start of this section should often be relaxed.
Example 7: Numerical Solution of Equations Example 8: Sequences of Heads and Tails

An important role of mathematics is to guide us
There is no formula for most equations f(z) = 0
in making decisions under uncertainty. This can of-
that need to be solved in real applications, so one
ten be done using probability theory, but often the
must use numerical approximations. A common ap-
most direct approach is simulation. To take a very
proach is the bisection method. If f(z) is continu-
simple example, suppose we flip a fair coin until we
ous, and one can find input values a and b with
get two heads in a row. How many flips should we
f(a) < 0 and f(b) > 0, then there is at least one
expect to take? If we actually carry out this ex-
root in between. (f(a) > 0 and f(b) < 0 is just as
periment many times, we find out what to expect.
good, and below we cover both cases by the condi-
Here is a algorithm to carry out the experiment one
tion f(a)f(b) < 0.) Try the midpoint c = (a + b)/2.
time. Rand(O,l) is a command for flipping a coin;
It is unlikely that f(c) = 0, but the sign of f(c) tells
the output 1 means heads, 0 means tails. The algo-
us which half of the interval [a, b] to look in further.
rithm could be run a thousand times inside a loop
Now iterate:
of a bigger algorithm, which could then analyze the
output data in various ways (take the average, the
variance, draw graphs, etc).
Input (none) buttons, approach (4) is available while approach

Algorithm (3) is not.
count +- headct + 0 Still other approaches are possible. Using the
repeat distributive law, we could instead evaluate
flip + Rand(O,l) [0 or 1, at random] a61 + abz + . . . + ab,.
count + count + 1
The direct approach to this, using the fact that mul-
if flip = 1
tiplications are completed before addition (on my
then headct + headct + I calculator), is
else headct + 0
endrepeat when headct = 2 a x bl + a x b2 + . . . + a x b, =
Output count which involves considerably more button-pushes.
[total flips to get 2 heads in a row] But many calculators, mine included, have a fea-
ture to shorten repeated multiplication by the same
Now, this algorithm violates our definition in factor: hit the x button twice. Thus the following
two ways. First, it is not determinate: actions are string of steps displays first abr, then abz, and so
not uniquely determined. Second, it is theoretically on:
possible that it wont terminate - we might get OS a x x bl = bs = b3 = . .
forever. Nonetheless, we certainly want to be able Now we want to add these up, but hitting + (or any
to study such algorithms. The hard part, actu- other operator on the main display) will cancel the
ally, is to get computers to perform such procedures, effect of x x. So instead we push M+, the memory
since computers really are determinate machines. In plus button, which does the addition in the hidden
other words, how can computers be made to produce memory register. Finally, at the end, we push MR
what appear to be random numbers? Fortunately, to remove memory:
there are good answers, using pseudorandom num-
a x x bl M+ b2 MS ... b, M+ MR (5)
ber generators.
Perhaps this sequence looks sufficiently odd that a
Example 9: A Calculator Exercise presentation in algorithmic language would help:
Except in Example 1 we have not said anything
about how our calculations are carried out; it could push a x x bl M+
be by hand, by calculator, or by computer. In fact, for k = 2 to n
what is easy to do depends on the device. In this push bk M+
example let us specifically consider hand calcula- endfor
tors, since one can hope that this product of mod-
push MR
ern technology can be made available to students
almost worldwide.
A count shows that method (5) takes the
Consider the problem: same number of button-pushes as the original ap-
evaluate a(bl + b2 + + b,). proach (3), and only one more than the best ap-
proach (4). So this problem provides a good exam-
How shall we do this? On my scientific calculator,
ple of how the issue of relative efficiency of algo-
which has parentheses buttons, I can do it exactly
rithms pertains to even very elementary mathemat-
in the order presented.
ics.
a x ( bl + bz + . + b, ) = (3) To close this section, let us emphasize that by al-
where each symbol now represents a button (except gorithms we do not mean computer programs. We
a, bl, etc., may represent many number buttons, and mean procedures for solving problems presented in a
represents repetition). However, we can save sufficiently precise form for careful analysis. While
time if we multiply by a on the right: we have written most of our algorithms in a style
which until recently has been associated only with
bl + 62 + .. + b, = x a = (4) computer programs, this is because that style is a
A + would do as well as the first =; the point good one for making key points precise. Our al-
is, the sum is computed as we go along, so once the gorithm descriptions cannot be input directly to
sum is finished, we can proceed to multiplication. any computer. They omit all sorts of information
One button-push is saved. Also, if you have only that a computer would need to know about (how is
a simple 4-function calculator without parentheses the data input and output, what type of variables
need to be declared, how much storage must be re- numbers have the same gcd? Yes, m - n and n
served?). Many computer scientists call this sort of have the same gcd as m and n, because anything
algorithm description pseudocode, because it is not that divides (evenly into) m and n divides m - n
real code for computers. But it is quite real for and n, and anything that divides m - n and n di-
the sort of communication that interests us here - vides m = (m-n) + n and n. And if subtracting n
between humans - and so we prefer to call it algo- from m once preserves the gcd, then subtracting as
rithmic language. many times as possible, leaving the remainder when
m is divided by n, also preserves the gtd. This is
2. What Is Algorithmics? the insight that leads to Euclids algorithm.
First, algorithmics does not mean performing The algorithm for Towers of Hanoi is also baaed
a lot of algorithms. Students worldwide have suf- on reducing to a smaller case. You. can solve the
fered too much rote repetition of mathematical al- game with n rings if you can solve the game with
gorithms over the years already. In the future, al- n - 1 rings, as shown in Fig. 2.
gorithms will be carried out more and more by ma- There are, of course, many other principles of al-
chines, or by person-machine combinations, so hand gorithm design, and whole university courses are de-
calculation except of the simplest sort should receive voted to it. Here well mention two more, top down
less emphasis. design and bottom up design. The former refers to
Algorithmics is the process of creating, under- outlining the big picture first, and then filling in the
standing, validating and comparing algorithms. In details of the parts later. The latter refers to start-
short, it is thinking about algorithms, not thinking ing with small pieces and putting them together
like algorithms. to do the whole job. While top down is generally
Here is another way to put this. The phrase al- the better approach for involved problems, both ap-
gorithmic mathematics has two meanings, tradi- proaches have their roles.
tional and contemporary [Maurer, 19841. The tradi- Algorithm design is more or less the same thing
tional meaning emphasizes carrying out algorithms, as problem-solving methodology. Since mathemat-
the contemporary emphasizes developing them and ics education is permeated with problem solving, al-
choosing intelligently among different algorithms for gorithm design is rightly an important component
the same task. of a modern mathematics education. Practice in de-
We now discuss the components of algorithmics sign not only makes people more successful at solv-
in more detail. It is standard to divide algorithmics ing problems, but also it results in algorithms that
into three parts, design, verification, and analysis. are easier to communicate to others and to verify.
Algorithm Design is the process of algorithm cre- Algorithm Verification is the process of confirm-
ation. There are some general principles of algo ing that algorithms solve the problems they claim
rithm design; it does not have to depend on un- to solve; in other words, proving algorithms correct.
teachable flashes of originality. Since loops are a primary aspect of algorithms, and
The most important idea, as in much of math- since a loop can be iterated any nonnegative integer
ematics, is to break a problem into pieces. If you number of times, mathematical induction is the key
can find a small building block that you underst,and, method of verification.
try to iterate on that block. To sum a sequence of Take Euclids algorithm (Example 2). Let P(k)
numbers, reduce to the case of summing two num- be the statement that, just before commencing the
bers; create a running sum and add one more num- kth pass of the repeat loop, gcd(num, denom) is
ber to it each time. To multiply two large num- the same as the gcd of the original m and n. That
bers (Example l), figure out a way to reduce it to P(k) is true for all k > 1 is easily proved by in-
many instances of multiplying two one-digit num- duction, using the fact that gcd(m, n) = gcd(n, r)
bers. To multiply two matrices (Example 3), first where r is the remainder when m is divided by n.
use the definition (2) to reduce this to many cases (We argued this fact informally when Example 2
of a real-number calculation, and then use iteration was introduced, and again somewhat differently five
to return this to single additions and single multi- paragraphs ago.) When the loop entrance condi-
plications. tion is tested for the last time, denom = 0, and
Sometimes one does not immediately see how to so clearly gcd(num, denom) = num, and num is the
reduce a large problem to small pieces. Then one value output. So by the induction, the output equals
tries to reduce it to slightly smaller pieces. What gcd(m, n) and the algorithm is valid. A proof of cor-
is the gcd of two large numbers n and n (Exam- rectness like this is called a proof by loop invariant;
ple 2)? Well, does some slightly smaller pair of the loop invariant is the statement you prove to be
correct each time you enter the loop. are both n x n, then there are n2 entries to com-
Or take the algorithm for Towers of Hanoi. We pute, and each entry requires n real-number multi-
may do induction on n, using the statement that plications and n - 1 additions. Therefore, the whole
any call of procedure H(n, r, s) correctly moves n algorithm takes n3 steps (if only multiplications are
rings from pole P to pole s. Since the very definition counted), and 2n3 - n2 (if additions and multiplica-
of H involves itself with n - 1 rings, induction is tions are counted). Or take Towers of Hanoi. The
easy to carry out.. In general, a recursive algorithm obvious thing to count is number of ring moves. It
immediately suggests an inductive proof. turns out that, if there are n rings, the algorithm
The specifics of how to do induction for algo- takes 2 - 1 moves. If t, is the number of moves
rithms is not the point here. The point is that in- with n rings, the recursive definition of procedure H
duction is the right tool. Mathematical induction, leads to the conditions
heretofore regarded in some quarters as a special-
t n+1 -- 2t, + 1, t1 = 1; (6)
ized method for proving certain formulas for sums,
must be viewed as a more central proof method in the unique solution of these conditions is t, = 2-1.
any curriculum that gives substantial emphasis to Calculations like these become valuable if the
algorithmics. number of steps appears large and one wonders
Algorithms need to be verified because more and whether the problem will be tractable with the com-
more our lives depend on them, and once they are puting equipment available. Suppose, for instance,
in place (say in our bank, to maintain our account that a problem requires n! steps when there are
records) they tend to get treated as black boxes. To n input data. (Brute force approaches to the fa-
be honest, algorithms used in the world at large are mous Traveling Salesperson Problem take this many
very complicated, too complicated for humans to steps, and the best exact methods known are in gen-
carry out detailed mathematical proofs of correct- eral not much better.) Then when n is merely 25,
ness; and machine verification of correctness is still a computer that could do a billion steps a second
in its childhood. Thus, empirical debugging tech- would still take 50 million years to solve the prob-
niques play a vital role. lem! In contrast, the same computer could play 25-
But mathematical verification should not be dis- ring Towers of Hanoi in only .003 seconds, and could
missed. First, big programs use many small build- compute the product of two 1000 x 1000 matrices in
ing blocks which can or have been verified. Second, a second.
the algorithms whose correctness you are primar- These efficiency calculations become even more
ily responsible for are the ones you create yourself, interesting when you have more than one algorithm
and knowledge of how to verify an algorithm can be for the same problem. Take Example 9 for comput-
helpful at the design stage. If you propose to in- ing a(bl + . . . + b,) on a hand calculator. The best
clude a loop in your algorithm, and you know that approach we discussed takes n+2+C button-pushes,
the way to validate it is with a loop invariant, you where C is the number of pushes needed to enter
will devise the loop invariant before you write the all of a,bl,... , b,; two others took n + 3 + C and
loop, and then you can write it to be sure that the the fourth approach took much longer. On a hand
loop invariant is preserved. calculator, each button-push takes time, so even a
Algorithm Analysis is the process of determining saving of one is significant. Furthermore, there are
how long an algorithm takes to run, and comparing lots of other, elementary problems where different
that run time to that of other algorithms for the calculator methods make a considerable difference.
same problem and to absolute standards for that Take the problem of evaluating a polynomial
problem. Run time is a rough way to put it, since
p(x) = anzn + an-l2 +-l + . . . + ao.
that suggests an actual machine (or person) to per-
form it, and different machines (and persons) will There are a great many multiplications involved, es-
perform differently on the same algorithm. Usually pecially if you dont have an exponential key. But
one picks some salient feature, say the number of there is another way to write this polynomial, best
real-number additions if addition is the main oper- understood in traditional notation if we use a nu-
ation in the algorithm under consideration, and de- merical example. If
termines the number of repetitions of this feature as
a function of the input size. This function is called p(x) = 4z4 + 3z3 + 2x2 + x + 8,
the complexity of the algorithm, or its efficiency.
then in nested form
Take, for instance, our algorithm for matrix mul-
tiplication (Example 3). If the two input matrices p(x) = x(2(2(4x + 3) + 2) + 1) + 8.
Calculating p(z) in this form takes many fewer steps ready pervasive in the world, and so informed citi-
(count them) and its easy to carry out even with a zens need to know what can be done by algorithms,
simple four-function, calculator and no memory. By how it is done, and how algorithms can be assessed.
the way, to make clear how both approaches work in Also, a fair number of people need to know how to
general, and to make the algorithms precise enough create algorithms.
to count the steps without confusion, the first thing Mathematically, this is an extrinsic justifica-
to do is figure out how to state them in algorithmic tion; algorithms are important, so students ought
language. to study them whether or not they are interesting
Or take the algorithm for constructing fi (Ex- mathematics or do good things for mathematics ed-
ample 4). It takes n - 1 right triangle constructions ucation. Fortunately, there is equally strong intrin-
to obtain fi. Noting this surely will inspire stu- sic justification.
dents to find a better way. First, introducing algorithmics in school raises
For Euclids algorithm, the number of steps de- fresh questions about old material and allows for
pends on the specific input, not just the size (num- greater student creativity. As Example 9 (calcu-
ber of digits) of the input. And in the Two Heads lator efficiency) shows, even basic arithmetic is no
algorithm (Example 8), there is no input at all, but longer cut and dried. Too many traditional curric-
the number of steps varies. In these cases one takes ula consisted of many computational courses where
several measures of the algorithms efficiency - best students were told the right methods, and a few
case, worst case and average case. Average case is proof courses (say, classical geometry) where they
especially important, but usually hard to analyze. were asked to be creative, but in a narrow theo-
The hardest problem is to compare an algorithm retical way. In contrast, each question of the sort
to an absolute standard. The complexity of a prob- devise an algorithm for . . . allows for many cor-
lem (as opposed to the complexity of an algorithm) rect answers (not all equally good). Even a student
is defined to be the number of steps needed by the who does not have a good theoretical grasp of the
best possible algorithm for the problem. Problem problem at hand may come up with a correct algo
complexity is the subject of much current research - rithm.
its hard to figure out the complexity if, as usual, Even incorrect algorithms can have worthy fea-
you dont know what the best algorithm is. For in- tures. They may involve good heuristics - imperfect
stance, it is known that Arabic multiplication and but insightful ideas that often lead to a reasonably
standard matrix multiplication are not the best al- good solution in a reasonable amount of time. Also,
gorithms for their problems, at least when n is quite an analysis of their flaws may be instructive and lead
large, but no one knows what the best algorithms to interesting class discussions. For instance, sup-
are or how fast they are. Nonetheless, progress has pose you want to pick a random set of two distinct
been made in finding bounds on problem complex- numbers from 1 to 10. Whats wrong with picking a
ity. And every once in a while the complexity of number i at random from 1 to 9 and then picking a
a problem can be determined completely. For in- number j at random from i+ 1 to lo? I once heard a
stance, it is not hard to show that the algorithm we businessman say, speaking at a college graduation,
gave for Towers of Hanoi is optimal. that in the outside world one learns from ones fail-
In closing this section, we note that there are ures. While there may not be much to learn from
other ways to analyze the goodness of an algorithm mistakes in traditional rote calculations, there is a
than speed. One can consider space complexity - great deal to learn from ones failures in devising
how much storage is needed. One can also con- algorithms.
sider numerical stability. For instance, in solving A second intrinsic reason for studying algorith-
a quadratic (Example S), if b > 0 and 4ac is very mics is: it can help students understand traditional
small compared to b2, then (b - s)/2a is practically mathematics better. You really have to understand
0, and roundoff error may swamp the computation. a procedure well in order to explain it to a com-
In this case it is better to set 22 to 2c/(b + s). Alge- puter, or to write it in algorithmic language, For
braically, the two formulas are equivalent, but, less instance, to understand Arabic multiplication well
roundoff error is introduced in the latter since b + s enough to describe it in algorithmic language (Ex-
is not near 0. ample l), you really have to understand place no-
tation and the distributive law. And its not just
3. Why Study Algorithmics? procedures that come to be understood better, but
We have already given our main reason: the use abstract concepts as well. For instance, the function
of sophisticated algorithms to solve problems is al- concept is concretized by seeing algorithms turn in-
puts into outputs. Real numbers are made more it is easy to show them that in the outside world
concrete as the student sees (Example 7) that suc- some very large problems must be solved where dif-
cessive rational approximations needed to compute ferences in algorithm efficiency are crucial. Sorting
them. Moreover, when students test their algorith- and searching (discussed later) provide good exam-
mic representations by running them on computers, ples; governments and large businesses must sort
they get instant feedback as to whether their con- and search enormous data sets.
structions are correct. The issue, then, is to keep the algorithmic exam-
There are also arguments against studying algo- ples up to date. This can be done if educators keep
rithmics in school. The basic argument goes: the informed about the latest research and the latest
sort of questions emphasized in algorithmics are al- technology.
ready outdated, or soon will be. For instance, al- Sometimes the exact opposite reason is proposed
gorithmics puts great emphasis on the relative ef- for not studying algorithmics. It is a theorem that
ficiency of algorithms. But if one approach to a there is no algorithm for determining which prob-
problem takes n2 steps and another takes n steps, lems are solvable by algorithms. (This is because the
the difference in actual seconds will be unnoticeable universal Turing machine cannot solve the halt-
for the values of n used in any classroom. Or, why ing problem.) So to emphasize algorithmics either
bother discussing methods for doing computations misleads them about what algorithms can do or cuts
on a four-function calculator when soon all calcu- them off from problems that have no algorithmic S(F
lators will be much more powerful? Indeed, whats lutions.
the point of talking about the merits of different But we do not propose that only algorithmic ap-
methods of polynomial evaluation, when on the now proaches to mathematics be studied. We only pro-
popular algebraic calculators, you can punch in pose that algorithmics receive much more attention
the definition of a function as a formula in 2, then than previously.
punch in a numerical value for 2 and finally just hit
the EVAL button? One doesnt need to know any 4. Suggestions For Implementation
method for breaking down the evaluation of the for- Two disclaimers: First, my knowledge of cur-
mula into small steps because the calculator does it ricula worldwide is limited, and so I speak mostly
all. from an American viewpoint. Second, in a paper
Generalizing, computing devices are getting this length, one can at most give illustrative exam-
more and more advanced in the sense that they ples and broad ideas of how to implement algorith-
can respond to higher and higher level commands. mics. For more detailed ideas, appropriate at least
When you think that, in the wings, there are main North America, see [NCTM 1989, Kenney 19911.
chines that will create proofs and create algorithms The suggestions below concern the primary and set- .
for solving problems, why do students need to be ondary levels except for a few brief remarks about
schooled in the ability to create algorithms them- the university level at the end.
selves? Look at traditional computations more closely.
I answer as follows. No doubt the level at which Basic arithmetic, computations with polynomials,
it will be appropriate to do algorithmic analysis will solutions of linear equations - such things are often
change over time. I really like to discuss different taken as routine and devoid of opportunity for fresh
methods of polynomial evaluation with my classes, thought. But from the viewpoint of algorithmics
but one day (perhaps soon) this may seem as out- there is plenty to think about. Students can dis-
dated to them as if I were to explain the theory be- cover traditional algorithms using design principles,
hind slide rules. But if we can draw any lesson from and discover alternative algorithms. While they are
the history of computing technology, humankind, unlikely to discover significantly faster algorithms,
including students, will always use technology to its they can be told ( or, at a higher level, shown) that
limits, and its most powerful use will always involve faster algorithms exist, and that best algorithms are
the interaction of human and machine. To pick a unknown.
simple example, I am not worried that the differ- Treat nontraditional computations related to
ence between an n2-step algorithm and an n-step classical questions. In every country students learn
algorithm will be lost on students. First, some stu- closed form solutions to certain sorts of equations,
dent always tries to run a recently learned algorithm but they dont always look closely at how to eval-
on data that is too large, and wonders out loud why uate those solutions accurately, or discuss methods
the machine sat spinning its wheels. Second, even if for approximately solving equations without solu-
most students stick to small textbook data sets, tion formulas. Students often learn to count per-
mutations and combinations, but they dont often lier, even four-function calculators are very helpful
consider how to efficiently list all of them of a certain in bringing to life algorithm design and efficiency
size, or generate a random one. In short, classical questions. With computers as well as calculators,
formulas that dont appear algorithmic raise alge one can start in the early years with such things as
rithmic issues. the language Logo and Turtle graphics, and move
Introduce some new topics. There are whole in later years to computer algebra systems.
fields of mathematics, with many applications, that Introduce algorithmic language. Whatever com-
have an algorithmic flavor and are not represented puting power is available, precise methods for de-
at all in many curricula. Many of these are grouped scribing algorithms are necessary if algorithms are
these days under the headings discrete mathematics, to be an object of study and not just something
operations research and theoretical computer sci- students perform. There is no standard algorithmic
ence. Here are a few examples, but at this point they language, and perhaps different sorts of languages
are little more than name-dropping, and one should are best for problems to be treated with different
refer to texts in these fields, such as [Hillier and sorts of machines (or by hand). Nonettieless, it is
Lieberman 1986, Manber 1989, Maurer and Ralston not hard to devise useful language constructs.
19911. Put more emphasis on mathematical induction.
Difference equations is the study of inductively We have indicated how induction is the main
defined sequences such as the step-count sequence method for validating algorithms. Actually, induc-
t, of Towers of Hanoi in Display (6). These include tion can be viewed more broadly, and as such is at
the traditional arithmetic and geometric sequences the foundation of algorithmics. There are inductive
and series - and much more. Computing terms in discovery techniques (reduce to the previous case, or
inductively defined sequences is immediately an al- build up from small cases to find a pattern), induc-
gorithmic question, and conversely, analyzing algo- tive definitions, as in Display (6), inductive algo-
rithms reduces to analyzing difference equations. rithm commands (loops and recursive procedures)
Graph theory, in the sense of networks, is full as well as inductive proofs.
of algorithmic questions. If a graph represents an Eliminate the schism between solving and com-
existing road network, how do you find the shortest puting. Traditionally there is pure mathematics and
routes between points (in distance, time, or what- applied mathematics. Pure mathematicians prove
ever)? If the network represents the possible links that solutions exists, and applied mathematicians
between cities is a telephone network planned for a figure out how to find them. In algorithmic math-
developing region, how do you decide which set of ematics, good computation methods are found si-
links will connect up the region at minimum cost? multaneously with showing that solutions exist. By
There are a variety of good (and not so good) algo- putting these two issues together right from the ear-
rithms for such problems, and many of these algo- liest years, we help to overcome what has sometimes
rithms are not hard for students to discover. been an unfortunate two-class system in mathemat-
Sorting and searching (e.g., alphabetizing and ics and science.
looking through an alphabetized list for a word) are A few words about the university level. Here
standard computer science examples that wouldnt the schism between pure and applied has been par-
traditionally be thought of as having any mathemat- ticularly acute. But it is breaking down. Many
ical content - clearly it is possible to sort and search, research mathematicians in pure fields are finding
so whats the problem? But once again, there are algorithmic questions interesting. Some algebraists,
lots of different methods, with various efficiencies, for instance, are now very interested in how classical
and various challenges to verify them and analyze objects in group theory can best be computed [Bee-
them. son 1990; Mines, Richman and Ruitenberg 19881.
Make computing power available to students. This could filter into the classroom. Even in calcu-
This is a tall order. No matter how rich the country, lus, some questions can be given a much more al-
there are always newer and more powerful devices gorithmic flavor than they have been. The rules of
one could want, and even in rich countries it may be differentiation, instead of simply being a set of rules,
a long time before there is one computer per student can be viewed as the parts of an algorithm that de-
in every class. But the point is, algorithmic ques- termines the derivative for any elementary function
tions take on much more life when students have (once elementary functions are given an inductive
what they regard as powerful computing aids, and definition!). The rules of integration can be viewed
then they discover they can still devise problems as part of an algorithm that determines the integral
that arent solved instantaneously. As discussed ear- of some elementary functions, and some discussion
can be added that integration is no longer an art, North Carolina School of Science and Mathematics
because there is an algorithm for determining ex- (Barrett, G. et al) [1991]: Contemporary Pre-
actly when a function can be integrated in closed calculus through Applications, Providence, RI:
form. Those university mathematicians who have Janson Publications.
gotten interested in algorithmic questions should be Peressini, A. et al. [1992]: Precalculus and Discrete
encouraged to share with their colleagues their ideas Mathematics, University of Chicago School
about how to introduce these new approaches in the Mathematics Project, Glenview, IL: Scott,
standard courses. Foresman.
Ralston, A. [1981]: Computer science, mathemat-
ics and the undergraduate curricula in both,
REFERENCES
Amer. Math. Monthly, 88, pp. 472-85.
Beeson, M. [1990]: Review of A Course in Con- Ralston, A. [1984]: Will discrete math surpass cal-
structive Algebra, Amer. Math. Monthly, 97 culus in importance? College Math. J., 15
(April), pp. 357-362. (See also Mines below.) (November) pp. 371-382 (Forum article and
Hirsch, C. and Zweng, M. [1985]: The Sec- responses).
ondary School Mathematics Curriculum, (1985 Ralston, A. and Young, G. S., (eds.) [1983]: The Fu-
NCTM Yearbook), Reston, VA: National ture of College Mathematics: Proceedings of a
Council of Teachers of Mathematics. Conference/ Workshop on the First Two Years
of College Mathematics, New York: Springer-
Hillier, F. and Lieberman, G. [1986]: Introduction
Verlag.
to Operations Research, 4th ed., Oakland, CA:
Holden-Day.
Kenney, M., (ed.) [1991]: Discrete Mathematics
Across the Curriculum, K-f2 (1991 NCTM
Yearbook), Reston, VA: National Council of
Teachers of Mathematics.
Malkevitch, J. et al. [1988]: For All Practical Pur-
pose, San Francisco: W. H. Freeman. Also
available as 26 videotaped TV programs.
Manber, U. [1989]: Introduction to Algorithms, A

Creative Approach, Reading, MA: Addison-
Wesley.
Maurer, S. [1984]: Two meanings of algorith-

mic mathematics, Mathematics Teacher, 77
(September) pp. 430-435.
Maurer, S. [1985]: The algorithmic way of life is

best, College Math. J., 16 (January) pp. 2-18
(Forum article and reply to responses).
Maurer, S. [1991]: Proofs and algorithms: a reply to

Gerstein, UME Trends, Vol. 3 (January) p. 8.
Maurer, S. and Ralston, A. [1991]: Discrete Algo-

rithmic Mathematics, Reading MA: Addison-
Wesley.
Mines, B., Richman, F. and Ruitenberg, W. [1988]:

A Course in Constructive Algebra, New York:
Springer-Verlag.
NCTM [1989]: C urriculum and Evaluation Stan-

dards for School Mathematics, Reston, VA:
National Council of Teachers of Mathematics.
ON THE MATHEMATICAL BASIS OF COMPUTER SCIENCE
Jacques Stern
I?quipede logique, Universitd Paris
et
Departement de Mathematiques et Informatique, l?cole Normale SupCrieure
France
It is now clear to an anybody that a working first burden of the theory to provide a suitable crite-
mathematician cannot ignore computers: as a con- rion for drawing a limit between what is computable
sequence, it is commonly admitted that students in (or eflective) and what is not. A simple way would
mathematics, and especially those who intend to be be to use the word computable for everything that
teachers in the field, have to be exposed to some can be processed on a real computer. Although this
high-level language (such as Pascal). Nevertheless, point of view is not completely meaningless, it re-
this is far from enough: the question of whether stu- mains rather vague and cannot be considered as a
dents in mathematics should be familiar with some genuine mathematical notion because of its lack of
parts of the theoretical foundations of computer sci- precision. Furthermore, this point of view is not
ence cannot be avoided because these topics are pre- even historically correct: a lot of outstanding work
cisely the parts of computer science close to mathe- connected with the subject of computation theory
matics and seem to be necessary in order to estab- was published before the first modern computer was
lish connections between both fields that go beyond built. For example, note the work of Turing [1936],
the ability of using the computing power of modern Post [1936] on computation theory itself, and also
machines. the work of McCulloch and Pitts [1943] on the mod-
In France, following this line of ideas, the study elling of neuron nets, from which the theory of au-
of algorithms and related topics has become, in tomata grew.
most universities, a significant part of the standard It is precisely the theory of automata that we we
curriculum leading to graduation in mathematics. propose to choose as a starting point. Many reasons
Also, an optional test in computer science has been can be put forward in order to justify such a choice.
offered for a few years in the well-established Con- The theory is simple, established on firm mathe-
tours dAgrCgation de Mathematiques, which is a matical grounds and provides various exercises in
kind of teaching Ph-D, passed by most of the programming: for example, one can simulate an au-
teachers for the age-group 17-22. tomaton in a high-level language like Pascal or dis-
The author has recently published a book en- cuss algorithms that compute the minimal automa-
titled Fondements Mathdmatiques de lInforma- ton. Also, the concept of non-determinism, which
tique[I990], which covers a large part of the re- is of utmost importance in theoretical computer sci-
quirements in computer science for undergraduate ence, can be quickly and naturally introduced in a
programs in mathematics. The aim of the present simple setting. Finally, the theory of automata has
contribution is precisely to present some general several applications: to text editors and compilers
ideas that grew during the process of writing up in particular; this is not a minor argument.
that book. These ideas are my personal views al-
Nevertheless, one can easily come to the con-
though I owe a great debt to many colleagues with
clusion that automata do not provide a satisfactory
whom I have had inspiring discussions.
model for real machines. This conclusion can be
Before going into greater detail, let me make one
reached by writing down simple languages that are
remark: Mastering some of the basic tools in com-
not accepted by a finite automaton but also through
puter science will not turn a mathematician into a
the convincing observation that a central feature of
computer scientist. Instead, it should help to de-
computers is completely wiped out, namely their
velop a different frame of mind, suitable to under-
ability to store data in a memory. We are thus
stand the specific features of comput,er science. This
back to our original problem of defining the notion
is most important for a mathematician because, as is
of computable and it is reasonable, at this point,
shown in other contributions in this book, these spe-
to require that this notion should be described us-
cific features will necessalily affect bot,h the teaching
ing various different techniques that come out to
and t,he practice of mathematics themselves.
be equivalent: this will ensure that a mathemat-
Around the notion of computation ical invariant has really been found and this will
Computation Theory is considered by many peo- make Churchs Thesis highly plausible. (Recall that
ple to be a very dull subject; nevertheless, it is the Churchs thesis states that the notion of machine-
51
..- -._ ..I II--

computable function and the mathematical family putation, one can explain how a basic cost can be
of recursive functions are identical). attached to the execution of each instruction, the
Four distinct approaches can be taken. overall cost (or complexity) being the sum of all
Adding a memory device to a finite au- basic costs. Thus, one can define the complexity
tomaton. This yields the definition of a Turing function of an algorithm which measures its cost in
machine. terms of the size of the data. Of course this com-
Directly modelling actual computers. This plexity depends on the abstract machine chosen but
can be done through the notion of a random ac- one can check that, when one machine is simulated
cess machine (cf. Cook and Reckhow [1973]) by another, the complexity functions are polynomi-
operated by a very simple language similar to ally related. This allows the definition of the class
machine code. P of polynomial time computations, which is a rea-
Defining a simple class of programs. For sonable candidate for modelling a class of problems
example one can define a restricted version of sometimes called feasible or tractable.
Pascal which uses only the integer type and the
Around the notion of algorithm
control sequences if . . . then . . else and while
. . . do. Now that we are equipped with a theoretical no
Defining the class of (partial) recursive tion of complexity, it is necessary to use it in con-
functions. This is a good opportunity to discrete situations. This can be done through a re-
cuss functional languages: recursive definitions view of various algorithms. This review is, by no
can be handled by using constructs that are ex- means, an exercise in programming style, even if
actly similar to those appearing in Lisp. correct programs have to be written at some point.
The proof that all these definitions are actually The emphasis should be on the design and analy-
equivalent is a source of very interesting observa- sis of algorithms, which are very closely connected.
tions. For example, the fact that the restricted ver- Of course, the rules of the game should be clearly
sion of Pascal can compute all recursive functions stated and discussed, especially the choice between
proves the well-known fact that the goto statement the two main notions of complexity that are in use:
can be dispensed with. It may be worthwhile to note worst-case analysis and average-case analysis. This
that replacing while . . . do by for only allows the choice depends on the underlying model: for exam-
computation of primitive recursive functions. Also, ple, average-case analysis is relevant when the prob-
the simulation of a random access machine by a Tur- ability of ill-behaved cases is small. In both cases,
ing machine is a good exercise that shows how to the analysis is combinatorial in character and quite
handle a sequential memory. often yields non-trivial recurrence relations. In or-
Once the notion of a computable function has der to handle these, some specific tools are needed,
been given a precise definition, it becomes possible like the statistics of permutations and distributions
to discuss decidability issues: By coding Turing ma- and the use of generating series (cf. Knuth [1973]).
chines and constructing a universal machine, it does Generally, such techniques (e.g. the use of singu-
not require much more effort to state correctly the lar points of the generating series) only allow an
halting problem and show that it is semidecidable asymptotic analysis and one may ask if this kind
but not decidable (which means that a machine can of information has any practical meaning: after all,
find positive answers in a finite computation time the size of the data are bounded by the computing
but cannot do the same both for positive and nega- environment! It turns out that the asymptotic anal-
tive answers). It is not clear that the study of gen- ysis is actually relevant: When a given algorithm
eral recursion theory should be pursued. Still, one runs in time O(n logn), for example, it is usually
may wish to present the semantics of recursive pro- true that the constant implicit in the 0 notation is
cedures and the fixed-point approach to programs rather small and that the asymptotic behaviour is
and develop the recursion-theoretic tools that are reached rather quickly.
needed, such as Kleenes theorem (which basically The students should also get used to performing
states that the name of a recursive function can be the analysis of the complexity of an algorithm with-
used within its own definition). out going back to the original definitions, based on
Then, one can have a discussion on whether or abstract models of computation. If the size of the
not the dichotomy decidable/undecidable is of prac- integers is bounded (which is often the case in prac-
tical significance. This is a way to introduce Com- tical situations), the complexity is roughly the num-
plexity Theory through the constraints of time. Go- ber of machine instructions performed during execu-
ing back to the various mathematical models of com- tion. This validates the use of the overall number of
Mathematical Basis of Computer Science 53
comparisons as a measure of complexity for sorting This makes it possible to recognize whether or not
algorithms. When large integers are involved, things a given integer n is prime by picking random values
become a bit more complicated: a convenient, way of a and testing the above equality. If sufficiently
is to multiply the number of instructions performed many tests are successful, n is declared to be prime.
by n2, where n is the number of digits of the inte- At a more elementary level, examples can be
gers used. This is to take into account the cost of taken from the following list (which is not exhaus-
multiplication as O(n2). tive):
Together with algorithms the specific data struc-
Sorting. This should include a comparison of
tures used in computer science should be discussed:
various algorithms and a discussion of quicksort,
stacks, files, trees, graphs etc. It should be stressed
as an illustration of the power of the divide and
that this point of view is quite different from the
conquer method.
one that was taken in the previous section: In com-
putation theory, we considered simulations involv- Searching, with an emphasis on the choice of
ing basic manipulations on data structures and we specific trees as data structures.
claimed that these manipulations were not costly, Pattern matching, because of the connection
becauce we were interested in the general notion of with automata.
polynomial time. In practical cases, a given polyno- Graph algorithms, for the nice interplay be-
mial time algorithm can be superior to another one tween discrete mathematics and computer sci-
and, very often, the choice of a good data structure ence.
may actually save a significant part. of the running
Graph algorithms can be a way to introduce
time.
HF-complete problems. Indead, one can observe
The choice of algorithms that can be reviewed is that computing shortest paths can be done in a
quite large and depends on the mathematical back- very efficient way whereas no polynomial time al-
ground of the students. At an advanced level, it is gorithm is known for many graph problems, such as
probably more rewarding to give examples that use the Hamiltonian path problem. This problem can be
mathematics in a non-trivial way, such as: described in very concrete terms as follows: Given
l The fast Fourier transform and its applica- a set of cities together with possible air connections
tion to fast multiplication of integers (in time between them, can one tour all the cities, visiting
0( n log n, log log n)). each city once and returning to ones starting point?
l Basic algorithms for computer algebra. In order to handle this problem, one can
This can be an opportunity to demonstrate the
l guess a plausible solution
use of a computer algebra system, like Maple or
Macsyma. l check its correctness (in polynomial time)
l The simplex algorithm for linear program- All problems that can be solved in such a non-
ming. deterministic manner are called n/P-problems, and
l Primality tests, at least probabilistic ones. an h/P-complete problem is an n/P-problem that
can be used as a subroutine in order to solve
Unfortunately, it is not possible to discuss these
all other NP-problems, with polynomially many
algorithms in detail. and we will only briefly com-
extra steps of computations. It is an open prob-
ment on the last example. As is well known, test-
lem (probably the most important problem in the-
ing primality by sieving requires a large amount of
oretical computer science) whether or not an n/P-
time and memory. In order to overcome this diffi-
complete problem can be solved through a polyno-
culty, one may try to use the mathematical prop-
mial time algorithm, and as a consequence, n/P-
erties of prime numbers. For example, it is known
complete problems are considered to be difficult:
that, whenever p is prime and a is not zero modulo
They can only be attacked by time-consuming tech-
p, the so-called Jacobi a is equal to a*.
symbol
0 niques such as backtracking.
This is not the case in general. More precisely, if
Of course, the class of AD-complete problems
n is not prime, at most one half of the possible os
can be given a formal definition through non-
satisfy the equality
deterministic Turing machines. Once this is done,
a one can prove Cooks Theorem (Cook [1971]), stat-
- = a* (mod n).
0n ing that the satisfiability problem for clauses of the
As was observed by Solovay and Strassen, the com- propositional calculus is MF-complete. More ex-
putations required to compute Jacobi symbols and amples of n/P-complete problems can be given (cf.
exponentials modulo n can be performed efficiently. Garey and Johnson [1979]), such as:
l The travelling salesman problem include at most one positive literal. The resolution
l The knapsack problem method is a way to derive a contradiction from a set
l The clique problem of clauses by making systematic use of the tautology
Finally, some indications can be given on how to
handle NP-complete problems and also on practi-
cal applications of these notions, through the use of As far as the predicate calculus is concerned,
one-way functions. it is almost compulsory to use a constructive ap-
proach based on Skolem functions and Herbrands
Around logic theorem. (Recall that Skolem functions ensure that,
Many authors now emphasize the role of logic in whenever a formula %a(~, yr , . . . , yn) holds, a pos-
the foundations of computer science. This is pre- sible solution z of this can be computed by a term
sumably because of the deep connection that exists f(Yl,..., yn).) Herbrands theorem states that, pro
between computer programs and proofs. This con- vided Skolem functions exist, any set of formulae
nection was already implicit in the section on com- from which no contradiction can be derived can be
putation theory: The undecidability phenomenon is realized in a model whose domain is the set of closed
closely related to Godels Incompleteness Theorems terms). This provides both completeness and com-
that show the extreme limits of deductive mathe- pactness by reduction to the propositional calculus.
matics. At this point, one should not avoid discussing un-
It is therefore necessary to include a thorough decidability issues again: even if one starts with a
introduction to logic in order to endow mathemati- finite set of formulas, one usually gets an infinite
cians with a synthetic view of computer science. But number of Herbrand clauses and therefore the Her-
it should be added that the interplay between math- brand procedure does not necessarily come to a stop.
ematical logic and computer science is such that In the above setting, the search for a more ef-
logic cannot be taught now as it it was before the ficient procedure leads to Robinsons unification-
advent of computers. This applies both to the for- resolution algorithm, and as in the case of the
mal presentation of syntactical objects and to the propositional calculus, one has to restrict oneself
development of the theory itself. to Horn clauses if one is really concerned with ef-
From the formal point of view, it is extremely ficiency. As is well known, such a restriction en-
helpful to follow computer scientists and to consider ables the use of backtracking and this is basically
formulas as trees and not only as strings of sym- the strategy of the Prolog language. The study of
bols, as was done classically. With this approach, Prolog offers a very interesting application of logic
a notion such as a free occurrence of a variable is in computer science. It shows that the views of ar-
given a clear, almost geometrical definition, which tificial intelligence can be turned into an actual pro-
was not the case when it was introduced through gramming methodology. Of course, it is clear that
a cumbersome recurrence. This can be quite im- Prolog is a programming language and not a thee-
portant considering the fact that syntax must be rem prover and that completeness is lost, as a con-
quickly understood by students who have not been sequence of various features of the actual language,
exposed to logic beforehand. as the lack of the so-called occur-check and the
For the same kind of reasons, students have to be use of the cut primitive. In order to show how
motivated as early as possible. Indeed, this can be the language works, simple programs can be writ-
done by discussing the aim of artificial intelligence: ten and discussed.
How to make correct inferences from a database of Now, Prolog is not the only example of appli-
known facts. This is meaningful even in the simple cation of logic to computer science and one can
framework of propositional calculus and the diffi- choose to give an exposition of program verification
culty of the problem can be understood by recalling through Hoares logic. Recall that this method is
that the satisfiability problem is h/P-complete. The based on cutting the execution path of a Pascal-like
search for solutions to the deduction problem that program into loop-free pieces. To each cutpoint A
are not brute search algorithms leads to the method is attached a formula +A, whose free variables are
of resolution, which can be made very efficient in the the actual variables of the program. Logic comes
particular case of Horn clauses through linear res- into the picture in proving that, if execution leads
olution. For the convenience of the reader, let us from A to B and if 4~ is true at A, with the current
recall that clauses are disjunctions of literals; liter- values of the variables, then 4~ is true at B, with
als are either positive, i.e. propositional variables or the resulting values of the variables. This is used
negative (negation of such variables). Horn clauses to show partial correctness of the program, which
Mathematical Basis of Computer Science 55
means that, if execution terminates, the final for- Conclusion

mula expresses that the result is as expected. Total In this short paper, we have tried to describe
correctness can be proved along the same lines by
what we consider as the mathematical basis of com-
using a well-founded relation and proving that loops puter science, to show how the chosen topics can be
decrease values of the variables with respect to this
organized and to motivate the choices that we have
well-ordering. made. Following the further developments of com-
Finally, another topic where logic and computer science, these contents will presumably have
puter science interact at the conceptual level is to be expanded or modified. For example, it may
the &calculus, considered as another approach to appear important to discuss boolean networks (to
program-correcness. Once again, it is based on the model VLSI) or to introduce tools for the study of
connection between algorithms and proofs. This relational databases. In any case, we feel that math-
time one is talking about formalized proofs within ematical tools for computer science will become a
the framework of intuistionistic logic (without use of part of any advanced curriculum in mathematics.
the middle-third) and about systems of rules using
the typed-&calculus, where a proof yields a term t, REFERENCES
which is, in a way, its algorithmic content (cf Kriv-
Aho, A.V., Hopcroft, J.E. and Ullman, J.D. [1983]:
ine and Parigot [1990]). In order to be more precise,
Data Structures and Algorithms, Reading,
let us recall that X-calculus builds terms from vari-
MA: Addison Wesley.
ables, through the following rules:
l if t and u are terms, then (tu) is also (application Aho, A.V., Sethi, R. and Ullman, J.D. [1986]: Com-
of t to u) pilers, Addison Wesley, Reading, Mass.
l if t is a term and t a variable, then Xz.t also Clocksin, W.F. and Mellish, C.S. [1981]: Program-
(abstraction) ming in Prolog, Berlin: Springer-Verlag.
The X-calculus can be considered as a kind of
Cook, S.A. [1971]: The complexity of theorem prov-
machine language, a term being turned into a so-
ing procedures, Proc. 3rd Annual Symposium
called normal form by reduction rules. In order to
on the Theory of Computing, 29-33.
program a function with integer arguments (for ex-
ample), one proves a formula (stating that, the re- Cook, S.A. and Reckhow, R.A. [1973]: Time
sult is an integer). This gives a term t and execution bounded random access machines, J. Com-
is just the reduction of the application of t to the puter and Systems Science 7, 354-375.
terms denoting the arguments. Because of the way Garey, M.R. and Johnson, D.S. [1979]: Computers
programming is performed, correctness is ensured. and Intractability, A Guide to the Theory of
Of course, the work on this type of programming NP-Completeness, San Francisco: Freeman.
strategy is only beginning and one should not con-
ceal that the resulting programming style is highly Hoare, C.A. [1969]: A n axiomatic basis of computer
inefficient at this stage. programming, Comm. ACM 12, 576-580.
Kleene, S.C. [1939]: G eneral recursive functions of
More on syntax
natural numbers, Math. Annalen, 112, 727-
Because of the organization of our paper around 742.
computation, algorithms and logic, we have not
discussed some quite interesting connections where Knuth, D.E. [1973]: The Art of Computer Program-
mathematics provides the necessary background. ming: vol 3: Sorting and Searching, Reading,
For example, we mentioned that logical formulas MA: Addison-Wesley.
can be considered as trees and the same is true of Krivine, J.L. and Parigot, M. [1990]: Programming
computer programs. Now both usually appear as with proofs, J. Inf. Process. Cybern., EIK 26,
strings of symbols. It is therefore very important 149-167.
to be able to recover the full tree structure from its
Manna, Z. [1964]: Mathematical Theory of Compu-
string version. This is a part of compilation, called
tation, New York: McGraw-Hill.
syntat analysis (cf Aho, Sethi, IJllman [1986]). It
turns out that the theory of context-free languages McCulloch, W.S. and Pitts, W. [1943]: A logical
is exactly the tool needed to perform syntax analysis calculus of the ideas immanent in nervous ac-
efficiently. tivity, Bull. Math. Biophysics, 5, 113-115.
Post, E. [1936]: Finite combinatory processes for-
mulation I, J. Symb. Logic, 1, 103-105.
56 Inffuence of Computers and Informatics on Mathematics and Its Teaching
Robinson, J.A. [1965]: A machine oriented logic

baaed on the resolution principle, J. ACM, 12,
23-41.
Rogers, H. Jr. [1967]: Theory of Recursive Func-
tions and Effective Computability, New York:
McGraw-Hill.
Sedgewick, R. .[1983]: Algorithms, Reading, MA:
Addison-Wesley.
Stern, J. [1990]: Fondements Mathematipues de
Ilnformatique, Paris: McGraw-Hill.
Wirth, N. [1986]: Algorithms and Data Structures,
Englewood Cliffs, NJ: Prentice-Hall.
THE EFFECT OF COMPUTERS ON THE SCHOOL MATHEMATICS CURRICULUM
Klaus-Dieter Graf
Freie Universitat Berlin, Germany
Rosemary Fraser
Leo H. Klingen
Helmholtz-Gymnasium, Bonn, Germany
Jan Stewart
Bernard Winkelmann
Universitgt Bielefeld, Germany
Introduction
Historical Sketch and Trends give a review of some effective and successful steps
The three traditional cultural techniques (Kul- and some reasonable trends in the pursuit of this
turtechniken), which play the most important role goal in school mathematics.
in our childrens education are reading, writing and In addition, many of the examples of this pa-
calculating. From the time of their definition per indicate that the technology is already a signifi-
(perhaps 1200 years ago; Alkuin, an adviser of cant factor in school classrooms, a factor that more
Charlemagne, mentioned them) the sets of methods than deserves its place. The contribution that it
establishing these techniques have undergone great
can make to the social and academic interactions is
changes and so did the subsets which were accessible vivid and, once experienced, always valued.
at school levels. In our times the largest expansion
Finally, just as children play out a wide range
occurred in calculating, which developed into a tech-
of roles in being part of the community they are in,
nique of solving problems formally with numbers,
symbols, graphics and words. On one side, this is so too can computers. Thus we ask the reader to
consider the computer as a member of the classroom
a result of extensive mat,hematica.l research, which
among other results brought about powerful algo- community, one that is able to contribute to the
days activities in an appropriate fashion.
rithms, easy to execute. On the other side this trend
was accelerated by the rise of powerful processors for Considerations and concrete suggestions for the
algorithms, namely computer systems together with use of computers in mathematics teaching depend
their scientific background, informatics (i.e. com- on knowledge about and experience with such in-
puter science). These aids make a variety of formal struments shared by teachers and mathematics edu-
problem-solving methods accessible for school math- cators. Fifteen years ago these people had access to
ematics and other subjects, which previously could computers mostly as programmers in numerically-
not be executed by students and pupils. Algorithms oriented languages. So computing power was mainly
form one important class of these methods. used in secondary math education for numerical al-
The development outlined above caused and still gorithms in the form of short Basic programs. Ten
has a significant impact on school mathematics ed- years ago, another step - but still in the algorithmic
ucation. At least three of the didactical dimensions spirit - was taken with Logo on various home com-
of the mathematics classroom are envolved: content, puters with its underlying philosophy of exploring
method and medium, to say nothing of the pupil - mathematics in specially designed microworlds and
teacher relationship. Control on these impacts can of learning mathematics by teaching it to the com-
only be gained by integrating and organising them puter; Logo also included the use of geometry and
into mathematics curriculum at all levels, since, as symbolic manipulations. Primary education was in-
A. Ralston [1990] points out . . only . . curricu- volved with these ideas, even kindergarten.
lum content can serve as a lever to change the en- The proliferat,ion of so-called standard software
tire mathematics education system. Comput,er use on personal computers in the last decade gave way
in mathematics education started as a very special to new considerations and experiments, especially
method with mostly special topics. Future co& with spreadsheets, programs for data representa-
puter use should be a standard method, applied in tion, statistical and numerical packages, dat,abases,
whole strands of subject matter. This article will CAD (Computer Aided Design)-soft,ware and com-
57
__ ._._ ..--..-_1__, -*----

puter algebra systems. But in the beginning such At the secondary levels we consider two main
software was not very user-friendly, and afterwards aspects which influence the goals and aims of math-
became too complex; the need soon became obvi- ematics education: the (mathematical) preparation
ous for special school adaptations which allowed of students for their lives and occupations, and the
easy specializations, employed mathematical nota- role of mathematics and its applications in society.
tion similar to that used at school, and used power- The students preparation for their lives and oc-
ful and helpful metaphors, so that even users with cupations starts in the first instance at school with
little training and only occasional practice (as is typ- its various disciplines. Since through the availabil-
ical of school users) could succesfully handle them. ity of computers, there are now strong tendencies
This led to the creation of general and didactical to introduce simulations into the school teaching of
software tools which sometimes also had a tutorial science, most notably in biology, or of introducing
component, thereby integrating some traditions of elements of statistics and data analysis into the mea-
computer-aided instruction (CAI). All these forms suring sciences and geography (cf. Winkelmann,
of using the computer came into being in sequence 1987), this is obviously a challenge to the teaching
but can now be found simultaneously in discussions of mathematics: Mathematics should elucidate the
about mathematics teaching. principles, possibilities and possible pitfalls of these
Even if suitable hardware and software are now methods; ad-hoc-explanations of such methods by
available for ordinary schools, several necessary in- the specific content-oriented disciplines are surely
gredients are still missing: Teacher training is far not appropriate for giving the student a coherent
from sufficient; hardware availability in most schools appreciation.
is still dictated by the needs of computer science and It is important to realize that routine calcula-
computer awareness courses and the concentration tions of all complexities will be done increasingly by
of machines in special locations prevents or makes ubiquitously available machines which must be con-
difficult the natural, selective use of software - e.g. a trolled at various levels by the users concerned. This
function plotter - during short episodes in the teach- requires more insight, more breadth, more ability
ing process. to check consistency, but fewer routine algorithms.
Such an emphasis belongs to the perennial goals of
Influences on the Goals and Aims of
mathematics teaching, of course, especially in the
Mathematics Teaching new math movement. But now there is really the
In elementary schools children meet basic pro- possibility of leaving out some of the drill because
cesses with patterns and numbers in the mathemat- technology can take over. Even an insight into the
ics classroom for the first time. There is a range of fundamentals of computers and their programs may
uses of technology that have proved positive and belong to the preparation for life. This can often be
stimulating in helping children to express them- shared with the other formal discipline, informat-
selves and to progress in a confident and enjoyable its/computer science, if it is implemented. It is hard
fashion. In particular these can help to discovery - to be more specific, since the determination of the
partly unconsciously - of the importance of underly- elementary and more advanced cultural techniques
ing structures as an aid to qualified communication which are needed by the future citizens presupposes
in language and problem solving. The computer is a futurist view of society which is notoriously hard
well-suited to setting up structures - this will be il- to specify.
lustrated in the examples that are discussed in detail As to preparation for vocations, for university
in the section on Illustrative Software below. (For studies, fundamental ideas and experiences in al-
a more comprehensive discussion of the influence of gebra, geometry and fractals, analysis, data analy-
computers on mathematics teaching, see the survey sis and statistics, simulation and chaos would now
by Fey, 1989.) seem to be necessary in different kinds of studies.
The emergence of multimedia technology means More specific preparat,ions for special vocations are
that our communication with computers and, in- again difficult to determine. For example, CAD
deed, amongst ourselves will employ words, pictures (Computer-Aided D esi g n which helps the construc-
and sound in equal partnership and will not be lim- tion of planar, spatial a.nd other objects on the com-
ited to a fixed sequential presentation. Although puter screen) is necessary for an increasing number
this article draws on the experience of using micro- of technical vocations, and this means the need for
computers in the classroom, it will also be relevant new and different qualifications in geometry; but
to the more sophisticated interactive video delivery what is exactly needed and how to build a curricu-
that is now available. lum to fulfil1 the needs of the trades remains unclear.
School Mathematics Curriculum 59
The same is also true for the other domains men- is certainly part of the world that the children will
tioned in this chapter; therefore, it is not laziness grow up in but one might feel it is not yet a part that
that the descriptions above are so general and un- children need to meet directly. Indeed, there are
specific. The general direction of necessary change concerns expressed in some countries that it might
can clearly be seen, but concrete decisions cannot be be positively harmful to allow the use of technology
built on scientific knowledge yet; we have to experi- before certain basic skills have been mastered.
ment and gather ideas, examples and proven results In the next section we shall look at some exam-
in concrete circumstances. ples of use under content headings although they
Mathematics education at school not only has also give rise to cross-curricula work. For ease of
the task of delivering to students the qualifications illustration we shall take Language Development,
asked for in vocations and daily life, but it should Early Science and Basic Mathematics as our main
also give insight into the role of mathematics in cul- categories. The decision not to limit the primary
.ture and society, into the fundamental possibilities school part of this article to mathematics is deliber-
for understanding and description offered by math- ate in view of the fact that most elementary school
ematics, and into connected assumptions and lim- teachers carry a responsibility for the major part of
itations. In this respect, on the one hand today a total curriculum. It is thus important that the
the greater part of the applications of mathematics use of computers be set in this context. However,
is transmitted by the computer and thereby influ- the Language and Science examples also have a rel-
enced in its character, as will be discussed in some evance to mathematical processes although this is
instance below, and on the other hand the computer not made explicit.
is fundamentally a mathematical machine and thus Before looking at the specific examples, it is nec-
its proliferation is a tremendous amplification of the essary to discuss the social situation that children
mathematization of our lives. find themselves in. Basically, there is a teacher to
whom they can turn and who organises their ac-
Primary School tivities during the day; there is a group of children
Computers and Calculators for Young that they work with, those they play with plus spe-
cial friends that they confide in. Thus children con-
Children
tribute to a whole range of interactions sometimes as
The greatest impact of computers on the learn- part of a large class, at other times with a smaller
ing of school mathematics has occurred in secondary group, often just to one other person and, finally,
school. However, we wish to begin by discussing the they must frequently work things out as an individ-
primary school curriculum for three reasons: ual. In short, the challenge that young children face
l a natural and basically positive attitude towards of being a member of the classroom community is
computers can only be achieved at this level. complex and demanding.
l since primary school determines a students life- Children need to develop good productive rela-
long attitude toward mathematics, we must use tionships and for this they need effective verbal and
all possible means - and the computer is one of nonverbal skills. Communication through body lan-
the most powerful of these - to create a positive guage and other nonverbal signals develop naturally
attitude during primary education. and requires no formal intervention. With the spo-
l it is necessary that teachers planning to use com- ken and written word the structure of the language,
puters in secondary school and even in universi- although not formally expressed, begins to be un-
ties understand what was done in primary school consciously absorbed and then actively used to build
and what the problems were there. new sentences and expressions. This somewhat sur-
The first major need to socialise with peer prising occurance indicates the importance of under-
groups and to share them arises when children move lying structures as an aid to communication. The
out of the home into regular contact with others possible role of the computer in this process was
at playschool or infant school. Here, also, serious mentioned above.
work starts in developing spoken and written lan- We shall analyse, albeit in a rather crude fash-
guage skills, learning about the world and meeting ion, the roles played out by teachers, children and
basic processes with patterns and numbers. Plenty computers in the examples that follow.
of play and creative opportunities are provided to Thus the focus of the following descriptions will
allow natural skills to flourish. be to consider the quality of the communication
How can technology help in this busy active in the classroom community and to identify struc-
happy environment of early childhood? Technology tures and roles that enhance the interactions be-
60 lnff uence of Computers and Informatics on Mathematics and Its Teaching
tween members of the community. ploration, this software is independent of country,

culture or age range. It offers a stimulus to explore
Illustrative Software language from many different angles and from many
different content areas. It can be used by groups of
1. Language Development
children just beginning to read or by groups study-
We describe here an extremely simple but power-
ing an authors style or even to consider a math-
ful program called DEVELOPING TRAY. It allows
ematical argument. In this activity the computer
teachers to type in pieces of text or poems they wish
plays the role of tasksetter and manager and pro
children to explore. The written material may be fa-
vides support to a rich and enjoyable learning expe-
miliar or unfamiliar; it may be related to a project
rience. It may be a task that an individual tackles,
they are studying or simply may have an interesting
but equally small groups and large groups can com-
language pattern or style. At first all the children
bine their talents to find the hidden text. Teachers
see on the screen is the punctuation - commas, full
and children can work together if the text is not
stops etc. Even this stimulates discussion - is it a
known to them. Thus there is the opportunity for
poem or a piece of prose? They make their decision
the teacher to join in the activity as a fellow pupil
and go to the scratchpad- a type of notebook built
rather than to share the role of a tasksetter with
into the program - to record their first predictions
the computer. The children find that they can use
about the nature of the text. Now they must buy
the structure of language and their previous expe-
letters. Every letter or letter pattern they choose to
rience in language to help solve the problem. DE-
appear in the text costs them points. Every correct
VELOPING TRAY stimulates communication and
word they type in or correct guess they make on
supports the strengthening of the use of structure
their note pad gains points. At first children tend
in language.
to buy letters arguing about those which are the
best value. As they see single letters and groups of
2. Early science
letters dotted about the screen, a pattern starts to
emerge. The following letters at the beginning of a The following description by Anthony Paddle de-
text: scribes work using EARLY SCIENCE. He considers
0-ce --o- a ---e a use of the computer that offers support to infor-
may suggest the familiar opening to a traditional mation structured in binary trees. A diagram such
story - once upon a time. The three letter word as this is shown below.
t-e may be guessed as the. The children can type
in any missing letters. Correct guesses are not only
accepted by the computer but are also placed in the
rest of the text. Thus one h typed in a correct
position places all the hs in the passage. Incor-
rectly placed letters simply vanish from the screen.
Thus the piece of writing is slowly revealed to the
class like developing film in a photographerss dark-
room (hence its name). There is great excitement
as a word or phrase is identified or as the range
Crcstcd
of words suggests the general content being writ- Tll
ten about. Prawns, shells, fish for example might
suggest a passage about the sea; it could however
be about working on a trawler or in a fishmongers
or part of a menu for a banquet. The children not
only have fun watching the text develop before them
but they also enjoy looking back at their notes on
their scratchpad to see whether their guesses were
right or wrong. It is not just an exercise in reading
and comprehension; it is about collaborating and
co-operating towards a common goal - and it is fas-
cinating for teachers to watch all the skills and in-
teractions generated.
As a supporting structure into which the teacher
or indeed the children may place any text for ex-
Several years ago, various versions of a computer only ten of them need to be answered to identify
game based on binary trees called ANIMAL ap- any one of the animals. In practice, reality tri-
peared in magazines, and books. A typical dialogue umphs over logic and keys cannot be designed that
with a computer running ANIMAL looks something well; nevertheless, only a small fraction of the ques-
like this: tions are relevant at one time. The remainder are
distracting clutter, and it is easy to become hope-
Are you thinking of an animal? YES lessly lost in a large printed key. ANIMAL avoids
Does it live in water? NO the problem by avoiding the clutter. Secondly, AN-
Does it fly? NO IMAL breaks down the highly abstract problem of
Does it walk on four legs? NO designing an identification key into simple, concrete
Does it go BOING? YES steps. To add the mouse to the key it is not nec-
Is it a KANGAROO? YES essary to think of all the attributes of mice or to
search for some essence of mouseness that will dis-
Are you thinking of an animal? tinguish mice from anything else. You are simply
asked to find one clear difference between a mouse
Only the boldface answers are typed by the and one other animal. The key-searching part of the
player; the questions themselves are stored in the program ensures that the other animal is the most
computers memory. Things become rather inter- similar one already in the key, so that the mouse is
esting when you think of an animal the computer inserted in the right place. This is not the only way
does not know about: of breaking down the key-building problem nor, if
the aim is to produce elegant finished keys, is it
Are you thinking of an animal? YES the best, Nonetheless it is easy and foolproof: If
Does it live in water? NO the individual questions work, the whole key will.
Does it fly? NO ANIMAL was not designed as a robust educa-
Does it walk on four legs? YES tional tool and suffers from a number of deficiencies.
Is it an elephant? (sic) NO It is not possible to correct any of the questions or
answers once they are entered - spelling mistakes
The animal you were thinking are permanent. Nor is there a facility for saving a
of was a? MOUSE key on tape or disk, or for printing it out on paper.
The language used by the program itself limits the
Please type in a question that use (Are you thinking of an animal? is a built-in
would distinguish an ELEPHANT question). It would be awkward to use it to classify
from a MOUSE? plants or rocks.
There are now several elaborations of this idea,
HAS IT GOT A TRUNK?
written for educational use, in which these problems
have been solved. THINK is one example which,
For a mouse the answer
while keeping the outward key format of ANIMAL,
would be? NO
has become a sophisticated tool for the creation,
In fact, the program starts each time knowing correction and searching of binary trees.
just one question and two animals. All the oth- A further development is offered by SEEK,
ers are added by the players in the same way as which comes in a package with THINK, several
the mouse. ANIMAL mimics a very simple learning ready-made keys and a program called INTREE for
process. typing in whole keys quickly. SEEK uses the com-
Clearly, the questions and animals are stored in puters graphics to display the questions in binary
the form of a key (or, equivalently, a binary tree), tree form. The questions appear in boxes and, de-
ANIMAL combines the functions of a key-searching pending on whether you give a Y or N answer, you
program and key-building one. Although intended are led down a branch to the left or right into an-
as a try to fool the computer game, it could be used other box containing the next question or the an-
quite seriously as an identification aid. As such it swer. At any stage you can move back up the tree
has definite advantages over a traditional printed and down another branch, so that the whole tree
key, especially for children. can be explored. SEEK makes the structure of the
The first advantage is that only relevant ques- information appear obvious. ANIMAL and its more
tions are displayed on the screen. A key containing direct descendants appear, by contrast to produce
1000 animals needs 999 questions but, in theory, questions from nowhere; they seem cleverer than
_. .i . . . . _.-_.
they are.
QUESTION YES NO
3. Using Key Handling Programs
1 Feel your powder?
Programs such as SEEK can be used in a surpris- Is it smooth or
ing number of ways with children. Obviously they floury? 2 3
can be used to.identify things if a suitable tree of
questions is already available. At the other extreme, 2 Put some in a
children can build their own tree from scratch, given teaspoon and
a set of rocks, twigs or kitchen powders, for instance. heat over a
There are also strategies that fall between these candle. Can you BAKING 4
two. If a class is planning to go pond-dipping, a key see lots of POWDER
to the commonest animals might be created in ad- steam?
vance, using information from books. New animals
can be added one at a time as they are found, pos- 3 Look through a
sibly over a long period. This may well be the best magnifying
approach with large, complex groups of things. The glass to see
initial skeleton tree can be designed so that its main if it is lumps
branches represent the major groups (nymphs, lar- or crystals. Is
vae, snails, worms, etc.) and the research involved it crystals? 5 6
in creating it can give focus to the childrens prepa-
rations for the first outing. 4 Put a drop of
In the classroom, identification exercises can iodine on your FLOUR ICING
provide very effective frameworks for practice of ob- powder. Does it go SUGAR
servational and experimental skills. A particularly blue/black?
good example is the POWDER tree supplied with
the SEEK/THINK package. On the surface it is 5 Put some in a
simply an identification key for common household teaspoon and
powders, such as sugar, salt, washing powder, flour heat over a
and baking powder. The questions, though, are not candle. Does it
just passive observational ones: Most of them ask smell like SUGAR 7
the children to do something to the powder and toffee?
watch its reaction. In the next column is part of
the key as produced by SEEK on a printer. 6 Put some in
There is no one way of classifying things. There water and shake.
may be generally accepted ways for groups like Do you get lots SOAP POLY-
plants, animals or rocks, but even these are subject of bubbles? CELL
to constant argument among scientists. If children
are to understand why things are classified the way 7 Put a teaspoon
they are, they need to explore and compare differ- of powder on a
ent ways. It is here that programs like SEEK display saucer and add WASH- SALT
their real value. By taking care of the overall organ- vinegar. Do you ING
isation of the tree, they let the children concentrate get bubbles? SODA
on close observation, comparison and the logical and
language aspects of choosing good questions. practical testing and research into their uses. The
Imagine that a group of children are trying to result may be quite a long list of differences, so the
identify some epsom salts using the POWDER tree. second stage is to decide on the best question to be
They will probably find that it is wrongly identified added to the tree.
as a salt. If they decide to extend the tree they Does it dissolve in water? is no good because
will be asked to find a question to distinguish the the both do.
two. This is no small challenge, finding the best Does it taste salty? may be ruled out on safety
question may take a lot of time, experimenting and grounds (someone may try to identify something
discussion. The first stage is to find out everything poisonous).
they can about the two substances by observing, Do you buy it at the chemist? requires prior
knowledge and would be impossible to answer if you Autocalc Options

really did not know what the powder was.
Does it have big crystals? does not have a clear Type of problem Subtraction
answer: It depends what you compare them with. Difficulty Level Own
Also the crystal sizes of both vary enormously. Top number 1 to 20
Does it have long, thin crystals? is better, as is Middle number 1 to 10
Does it turn into white cake when you heat it over Bottom number 0 to 20
a candle? Delay time 2 seconds
Some of these problems are quite subtle, and This option setting provides simple subtraction
children are unlikely to spot them until they try the problems for young students.
bad questions in a complete tree. Fortunately, all Imagine a class of children working on the ways
the more recent programs let you prune and repair in which they add 9 to numbers. The computer is
a tree without having to rewrite the whole thing; set to produce problems where the number is gener-
so children can learn from their mistakes and cor- ated between 0 and 99, the second number is fixed at
rect them with a minimum of frustration. A good 9 and the time delay of 3 seconds before the answer
way of identifying problems and sharing insights is is given has been set. Fifteen problems appear one
to encourage groups of children to test each others after the other and the children attempt to calculate
trees. the answer before it is displayed by the computer.
To simulate the experience complete the following
problems as quickly as you can:-
4. Mathematics
28 90 32 77 88 79 37 66
~+J+s+s+s+s+9+9+
AUTOCALC is another example of a simple
program that promotes considerable discussion and
sharing of processes. It enables children to articu-
late their own methods and ideas and haa proved Probably after the first try at this task the chil-
an extremely valuable way to build their confidence dren will feel that they might be able to get the
in their mathematical abilities. The children are answers in under 2 seconds so that they can have
challenged by the program to try out their men- anoiher go with reduced time delay. Some might
tal arithmetic skills and to review and compare the even like to go at producing an answer in 1 second!
range of possible processes. A large screen is needed After this activity the children are asked to say ex-
at the front of the classroom. The screen presents actly how they got the answers. The following list
the problems in the following format: of methods was the result of a class of ten year olds
sharing their ideas:
44 1. Helen decided to add one to the tens and then

take one away from the units.
+ 29
2. Jonathan was happy to count on his fingers but
--
didnt always have time.
3. Susan added 3 three times.
After a delay the computer then supplies the an-
4. Jo subtracted 1 and added 10.
swer to the calculation
5. Anne worked out how many to the next 10s.
This is then subtracted from 9 and the remainder
44 added
+ 22 e.g. 78 + 9 = 80 + 7 = 87
73 6. Simon added 2 four times, then 1.
7. Jane used different mathematics for different
The mode of the program is to generate such problems.
problems by selecting random numbers according to 8. Michael just knew the answers!
the parameters set at the beginning, using a chosen
operation and displaying the answers after a chosen The children greatly enjoy sharing their methods
time delay. The option screen used for defining the and trying out each others ideas. They are also en-
type of problem to be set is shown below: couraged to use calculators - various tasks are given
_ _ . . ..___. -. . . - . .._-_-_ - -_-__ - --- .-UI

that begin t,o expand their understanding of num-

ber and to encourage them to feel confident enough
to move int,o estimating outcoming numbers as well.
AIITOCALC can be run in a mode where the chal-
lenge is not to do the calculation but to spot ev-
ery time an incorrect answer is displayed. (This is
called the oops.mode.) The skills needed to suc-
ceed here are now dependent on having a good grasp
of number bonds and relationships. Another excel-
I;
P.S.yd CJOH RV?rO.ylhhG q C-
lent activity is to ask the children how many prob-
lems they can make combining a number between
O-99 and a number between O-9 that has the an-
swer of 5; e.g. 13 - 8 = 5; 1 x 5 = 5 etc. After
exploring the problem in groups, the children offer
their solution. This brings out some fundamental
mathematical processes - classifying sets within the
solution set - setting the initial conditions in order
to limit the solution set to a finite set and many
others. The children express these ideas in their
own language and, of course, they are not yet aware Critics might say that the activities promoted
of the generalisation of such ideas. However it is with AUTOCALC are not valuable because they
at this point that we become aware that this simple are dealing with numbers out of context to any real
computer program has given the children a stimulus problem. However we hope that the examples here,
that has caused them to become t,rue mathemati- which are only a minute part of the range of possi-
cians. In sharing their mathematical processes and bilities, show children becoming aware of their own
in valuing each others ideas they will build up con- power and thought processes and also taking over a
fidence in their own abilities to offer something to range of teacher roles at various stages. Feedback
the subject. In this way we can begin to remove the to the teacher of the childrens reasoning and the
fear that so many people leave school with in regard way in which they articulate this is a major contri-
to their mathematical abilities. A final stage 69 the bution of AUTOCALC.
discussion of the 5s problem is to watch the com- A few years ago Michael Girling (Her Majestys
puter doing the same problem and to write report Inspector) suggested that a definition for numeracy
to it,s parent,s on its performance. As the comput.er might be appropriate use of an electronic calcula-
applies an ext,remely simple algorithm (it just keeps tor. What number sense would one need in order
randomly generating problems but only displaying to qualify?
t,hose that give the result of 5), its performance is We suggest:
certainly open to criticism. Here are some of the 1. Instant command of single digit arithmetic
childrens reports:
2. Command of basic multiplication facts
3. Skill in estimation
4. Capacity to spot errors
5. Capacity to select which operations are appro-
priate in any problem
With the exception of 5 all these points are
strengthened by the activities possible with AUTO-
CALC.
Concluding Remarks
This section has taken just a few examples of
simple software to illustrate how computers can
have a stimulating and refreshing relationship with
children. We are keen that the computer becomes
an accepted assistant and friend of both teach-
ers and children.The use of Logo, data banks and
word processing, have not been discussed here as of the proper conditions under which the facts re-
many books and articles are available to the reader main true (the domain of validity), etc. As a con-
on these topics. Such languages and systems can crete example, let us look at the calculus (analy-
be employed to stimulate discussion and exitement sis). Phenomena are: The graphs of functions, say
such as is described here. However, they can make of f(z) = zsin l/z, the fact that sinz/t tends to 1
considerable demands on the users and we would as I tends to zero, the divisibility of 2 - 1 by z - 1
recommend that subsets of such systems are used and the form of the divisors, the formulas for the
to start with. Slow progress is being made with im- derivatives of elementary functions, the linearity of
plementing a curriculum that make effective use of the integral, or the shape of solutions to a specific
computers and calculators. This is due to the fact initial value problem for a differential equation.
that there is not as yet a great deal of curriculum To the domain of theories, there belongs the def-
support materials to introduce the range of learning inition and fundamental properties of the limit, the
activities that simple or complex computer software completeness of the real numbers, the definition of
can support. However, this material will gradually the integral, the limits of validity of theorems, and
emerge and there is certainly enough available to explanations of facts by arguments.
any enterprising school to offer children the advan- It is interesting, that there may be different pos-
tages of a computer in their classroom. sible theories, for example, Euclidean or Cartesian
Any school able to equip each classroom with geometry, with formalist or constructivist founda-
a single microcomputer would gain experience and tions. Or, in the case of analysis there are differ-
confidence within a matter of months rather than ent possible non-equivalent theories, the classical
years. Add to this provision a small laboratory for 6 - h-theory, non- standard analysis and different
word processing etc. together with a collaborative constructivist approaches. But all those different
staff exploring possibilities together and the scene theories explain - in different ways - the same phe-
will be set for an exiting time for children in such a nomena. And all the concrete applications of geom-
school. etry or calculus only rely on the phenomena, not on
the underlying theories. In a similar way, comput-
Secondary School ers and mathematical software work exclusively in
the realm of the phenomena; they can only exhibit
Phenomena, Theories,
phenomena. And they are able to show the phenom-
Experimental Mathematics ena even to students who have not yet mastered the
In mathematical knowledge one can differentiate theory.
between facts on the one hand and the insight into This is the point in our argument: In a mathe-
their necessity and their connections on the other matics class using mathematical software, students
hand, or between phenomena and theories. This will get to see and know a lot of mathematical phe-
distinction becomes clear, for example, in the do- nomena. The mathematical theory then has to ex-
main of the geometry of triangles: Examples of phe- plain these phenomena; thus mathematics shifts in
nomena are the observable facts, such as that the the direction of a science which orders, describes and
three angel bisectors meet in one point and simi- makes understandable facts that are already known
larly for t .ie perpendicular bisectors, that the sum and obvious even without explanation. This is in
of the in] er angles equals 180 degrees, that two tri- sharp contr&st to classical teaching methodology, es-
angles w rich have the two sides and the enclosed pecially in such domains where it was hard to ap-
angle eqi al have all other measurable parts equal, proach the phenomena without theory or advanced
the form [la of Pythagoras, etc. Most classical the- technology.
orems of school geometry belong here, but so also Here is an example. In the study of functions
do more qualitative facts such as: If two sides are and their transformations, traditional teaching de-
fixed in 1 !ngth, then the third side gets longer if the duced behaviour mostly from theory, since the ac-
enclosed angle is made bigger (up to 180 degrees). tual plotting of function graphs by hand was far too
There is row special software such as The Geomet- expensive, in terms of time and labour, in order to
ric Suppl ser or Cabri Gdometre which helps to find make students see the facts, for example, the graph-
such fat s by giving assistance in the making and ical translations connected with the transformation
systemat c variation of geometrical constructions. f(z) + f(z + u). With the help of a function plot-
In tl e domain of theory there is the logical or- ter they may observe those transformations, first
dering ,f facts (local and global), the insight into connected with a concrete f and a, then system-
the necessity of observed facts, the determination atically explored with free chosen examples, and in
between also formulated as hypothesis and verified teachware, which allow some unconventional ac-
by arguments. In this way, the temporal order of tivities which are closely related to the curriculum
phenomena and theories reverses, and gets closer to for grades 7 and 8.
the usual habits of mathematics as a research ac- The elementary didactical philosophy is that
tivity. Of course, such an approach has often been there should be two levels of action in geometry
used with mathematical content where exploration classes, when using a computer: On one level the
of phenomena was cheaper. pupils should learn the constructions manually with
The didactical paradigm just described has of- ruler and compass, as usual. On another level they
ten been referred to as experimental mathemat- improve their competence with these constructions
ics, but it has to. be stressed that theory is an in- by solving geometrical and applied problems with
dispensable part of it in order to be mathematics. graphics procedures on the screen which they per-
Just playing around with a function plotter does ceive as efficient and comfortable tools. In par-
not necessarily lead to insight. You normally need ticular, this use of computer graphics in the early
hints, ideas, hypotheses, questions in order to see years of secondary school has proved useful in three
something and get involved. (See Goldenberg, 1988 modes:
for more specific considerations and examples.) As Using procedures for ruler-and-compass con-
a counterexample, using fractal generating software structions which have already been understood
may give spectacular pictures of great esthetic value as building blocks for more complex construc-
but, if you stop at the phenomena, you wont get at tions without the need to repeat the elementary
mathematics with such software. You need at least constructions again and again.
general concepts such as self-similarity or symmetry, Using procedures for constructions in ways
which are also needed for the better understanding which cannot be realized with ruler and com-
and appreciation of the beauty of the pictures in- pass.
volved. Using procedures for large and technically dif-
ficult constructions, which demand many itera-
Software for Secondary School tions of elementary constructions.
Mat hemat its The Geometric Supposer fits in mode 1. We now
discuss two other s&ware packages, SYMMETRIC
We shall discuss this for thrcr content areas
TURTLES and KALEIDOSCOPE, which illustrate
Geometry, Functions and Data Analysis.
modes 2 and 3.
Geometry SYMMETRIC TURTLES (Graf, 1988)

Two software packages for geometry education It is well known that Logos turtle graphics can
were mentioned above: The Geometric Supposer help at the beginning of geometry education. As a
and Cabri Geomktre which allow constructions of tool which provides an extension of the ruler and
most of the problems of Euclidean plane geome- compass a running turtle has been developed.
try, A so-called draft mode allows the exploration of This follows the concept of Abelsons dynaturtle
consequences of moving one point in a figure while [Abelson and di Sessa, 19851, but without inertia.
keeping its connections to other points (see Fig. 1). To some extent you can use it like a pencil, con-
Descriptions and examples are given in Schumann trolled with keys.
[1990]. Here we shall describe two other pieces of Keys 1, 2 . . 9 put it in slow or faster forward mo-
tion on a straight line, key 0 stops it. Z or N lets the
turtle draw or not draw when moving. A, S, D, F
effect small (5 degrees) or larger (15 degrees) left or
right turns of the stopped or moving turtle. Q marks
the position of the turtle on the screen and deter-
mines a number for this point. This point can be
reached again via keys K or P. K turns the turtle in
its actual position heading for another point. This
corresponds to putting a ruler through two points.
P puts the turtle on an already marked and named
point. And so on.
L This running turtle allows construction of many
Figure 1 figures of interest in plane geometry. Besides this
turtle there are two more turtles, an axial symmet- than from the final picture. You see that a straight
ric turtle and a central symmetry turtle. They line remains a straight line, you see how the direc-
are controlled in the same way as the standard run- tion changes under axial symmetry and how it re-
ning turtle. But they do not only draw the figure mains the same under central symmetry. You also
controlled but also the figures symmetric image in see that a straight line and its picture are parallel
a different colour relating to the I- or y-axis or the under central symmetry, but have different direc-
centre point of the screen. This happens simultane- tions, etc. Figure 2 contains some examples. Unfor-
ously and pointwise. This mode of construction for tunately, the dynamic quality of the turtles can-
symmetries helps the user to recognise the proper- not be seen from these figures.
ties of the mappings immediately and more easily Figure 3 shows how the following question can
be examined: What happens when reflecting a tri-
angle in different positions relative to the axis of
symmetry or a point?
Figure 4 gives a systematic answer to the ques-
tion, How can quadrilaterals be generated by re-
flecting triangles?
Figure 2 Figure 4
First, it is convenient to choose a side of a trian-
gle as an axis of symmetry. Then with the turtle you
get a kite. The special case of an isosceles triangle
occurs if the angle adjacent to the axis is 90 degrees.
If this angle is greater than 90 degrees, then you get
a quadrilateral which is not convex. You can also
get a rhombus and square by starting with special
triangles. But you never get a general rectangle or
a parallelogram or a trapezoid. The central sym-
metry turtle, however, applied on the centre of a
side of a triangle produces a parallelogram imme-
diately. This is an exciting discovery. The choice
of this special point of reflection is suggested by the
experiments shown in Figure 3. Again, no trapezoid
occurs. This fact can result in geometrical discus-
sions. More details about these tools are given in
Graf [1988]. Some reactions of teachers and student
teachers to this kind of teachware and some experi-
ences in classes are also reported there.
Figure 3
KALEIDOSCOPE
In a paper by Graf and Hodgson [1990] it is
shown that the kaleidoscope can be a window to

some geometric concepts. These are elementary and
rich at the same time. They are rich in the sense
that they offer not only a mathematical model of
the kaleidoscope but also models for other worlds
like planes to be tiled or even like a fictional kalei-
doscope.
From a methodological point of view the mathe-
matical problems connected with kaleidoscopes can
be worked on at the following five levels:
1. Looking through the real kaleidoscope.
2. Reducing the kaleidoscope to a model with two
or more real mirrors placed on a sheet of paper
containing some figure. Figure 5
3. Abstracting the mirrors and their reflections on another mathematical phenomenon: Coverings
straight lines (axial reflections) constructed with of the plane. In the case of three equal mirrors
ruler and compass. or sides you end up with a perfect tiling of
4. Transferring these constructions to a computer the plane. Figure 6 shows the growing of such
graphics display. a tiling. Obviously a discussion will arise from
5. Using formal methods to describe the phenom- this about good and bad tiles.
ena (and PROVE theorems!!!), for example,
those of analytic geometry and linear algebra.
Here we can only give a few glimpses on the soft- ,,-7;-.
ware for simulating kaleidoscopic phenomena on the ,7. F-y7 -I +-
,- ;;yty : A, ,
computer and examples of patterns that can thus be -I: _ T- \,.---A , -,, ,, ,;-y
, , , , , ,v, T ,-
produced. -L-- \I, ,,-, , ,, i II ,,,L , ,
;-; I , ,y
a) Two-mirror kaleidoscopes: The main menu of- - .-
.
fers a choice of four different types of kaleido-
scope. Mode 1 leads into a dialogue about form-
ing a kaleidoscope with an arbitrary angle. The
user gives the positions of the axes and then the
position of the object to be reflected. The com-
puter then displays the two axes and the object.
It then constructs and displays one reflection af-
ter the other until the pattern is complete. This
can be done with a pause after each image or in
an automatic mode. Mode 2 allows one to se-
lect a kaleidoscope with angle 45, 60, 72, 90 or
120 degrees and then proceeds as above. Figure Figure 6
5 shows some steps in the development of a 60 c) Fictional kaleidoscopes: The examples given so
degree pattern and the same process for a 70 de- far considered the transfer from real kaleid-
gree pattern. One of the many questions which scopes to mathematical models, combining ax-
will arise after such experiments is: How many ial reflections and varying the type of kaleid-
reflections are there before the pattern begins to scope. Why not vary the mathematical model
repeat itself? Circular symmetry can be discov- and forget about reality? Central reflections
ered and discussed after such experiments. (half-turns) will give us a model of some fic-
b) Polycentral kaleidoscopes: These are built from tional kaleidoscopes having no physical counter-
a greater number of mirrors and thus produce parts. One case is to look at a triangle again, de-
groups of images around several centres spread- termined by three centers of reflection, and see
ing in all directions. Forgetting about special what happens after repeated reflections. We can
objects between the mirrors and just regarding get a pattern extending throughout the plane
the reflections of the triangles or quadrilaterals (see Figure 7), leaving some blank spaces.
and so on shaped by the mirrors you discover
A new situation occurs if we start reflecting a tri-

angle not about its corners but about the midpoints
of its sides, and go on reflecting the images about
these midpoints. A new tiling of the plane devel-
ops and - what is really surprising for the beginner
- this works well with any triangle. Next you can
turn to quadrilaterals and again you discover imme-
diately that a perfect tiling can be completed with
any quadrilateral, even a non-convex one (see Fig-
ures 8 and 9). So the fictional kaleidoscope brings
you back to a real problem and the search for its
correct mathematical solution.
Figure 9
We conclude this section about geometry with
some general remarks. There should always be a
very careful examination of the advantages for learn-
ing before the computer is used in some field of
mathematical education. There is no use in trans-
ferring manual or mental activities (like construc-
tions with ruler and compass) to the computer un-
less this brings about more efficiency in learning.
Another good reason for using the computer may
exist if the computer allows activities which the stu-
dents cannot achieve with their hands or brains.
Then the computer acts like an additional tool,
increasing the traditional abilities of the students.
SYMMETRIC TURTLES and KALEIDOSCOPE
are good examples of such tools. They allqw
0 additional help in exploring mathematical prob-
lems.
. a great variety of investigations with little effort
. easy experimentation
. the viewing of a very broad spectrum of complex

Figure 7 geometrical constructions which turn up when
studying reflections of complicated figures
- l doing manually impossible constructions like the
pointwise simultaneous construction of two or
more figures
. introducing and using simple methods of CAD
(computer-assisted design), a technique which
has replaced manual technical drawing to a con-
siderable extent.
I
Functions
Function plotting software in acceptable qual-

ity for use in schools is now available for nearly
all modern microcomputers, and there are now sev-
era1 sources of didactical material describing teach-
ing units, giving hints and providing exercises which
can be used by normal mathematics teachers. The
Figure 8 general idea of a function plotter, to plot the graph
corresponding to a user given function, can also be
inverted, namely to plot the graph and let the user
look for the term. This is realized in the Funktio- to the needs of the students. A specific option lets
nen raten (looking for the formula) part in G~aphiz. the program be used by two partners (individuals,
For example, the program plots the graph of a groups): the first gives the term and the other has
function - say f(z) = 22 - 3 - but does not show to guess it from its graph.
the term (Fig. 10). The user has to make use of the The simple idea of the program gains its moti-
information given in the graph to guess the func- vational and challenging character from the use of
tion term and put it in. The computer reacts by a sophisticated function plotter, which comes close
plotting (in another color, if available) the graph to the accustomed appearance of terms and graphs,
corresponding to the users term in the same coor- and from its deliberate generosity to an inexperi-
dinate system. If the user has not got the correct enced user. It is simple to use. The user is not
solution (Fig. ll), h e or she can now see the dif- penalized for wrong answers. And it has adequate
ference between the original and the guessed graph error control, not through comparing the users term
and use this information to debug, that is, to cor- with a predefined list of possible right terms, but
rect any mistake. As many tries as desired can be by numerically comparing the graphs with a certain
made. It is also possible to wipe out the screen and tolerance. So the software aims really to help users
see only the original function. to evolve and debug their knowledge about elemen-
tary functions and their standard transformations.
The program is to be used mainly by individuals
or small groups, in a wide variety of levels, grades 7
to 12 and up. It may be used for drill and practice,
..............................
...... P-f...:. ... .f ..:....: .... i .... . and, of course, for remedial work.
: : : : : : I : f : : : :
. . . . . . . . . . . . . . . . . . . . .._. .__. ;.4+...;.../; . . ..I . . ..I . . . . . . . . . :,..
: : : : : , ./. . . : :
: : : : :
Data Analysis
Statistical education - as mathematics educa-
tion in general - often has to cope with the problem
that, in order to solve real problems, the necessary
. . . .. . . . .. . . .. . . . ++...: . . . .. . . .. . . . .. . . . .. . :.._
techniques are taught and, in consequence, also un-
f cx>=
derstood by students in isolation; their proper con-
CT3
ditions of application, their region of validity, their
limits are perhaps theoretically known, but seldom
Figure 10 part of active knowledge. In order to overcome such
limited understanding, one method is to confront
students with problems connected to themselves, so
that they dont take the methods as neutral, but
of real importance. One of the goals of the soft-
ware Times is just to give students some real data,
connected to themselves, in order to analyze and to
draw conclusions from the data and thereby about
themselves.
The software allows experiments with reaction
times: The computer produces a specific signal and
one of the students has to react in a specific way,
for example, by pressing a specified button, and the
computer measures the reaction time. The process
f cx>=x-3 c-?-3*
repeats, and the data are stored into a file bearing
the name of the student. Another student does the
Figure 11
same procedure, and the data is compared. Which
The functions plotted by the program are of- student is better? Is the arithmetical average a fair
fered in different sets, organized according to dif- arbiter or is the median better? How should one
ficulty and type of function (linear, quadratic, cu- judge extreme values? The program offers several
bic, trigonometric, exponential, using absolute val- methods of comparing data, including some well
ues, etc.). The sets can be changed or augmented known statistical techniques. It calculates diverse
with a simple text editor by the teacher, according quantities such as averages, variances, the plot of
one distribution of values against the other It does continuous algorithmic strand which forms 15% of
&Q-plots, displays the data as time series, etc. In the curriculum in mathematics education during the
defending their results, the students hopefully learn nine years of German grammar school. (We begin
to judge cautiously, to see the techniques as helpful at year five since we do not consider the four years of
but normally not decisive tools, and the necessity of primary education.) The following list shows typical
properly interpreting the data rather than automat- algorithms and also their related subjects.
ically drawing conclusions after a routinely applied 5 Relations between the fundamental arithmetical
test. operations
Transformation between numbers with different
General Tools and Methods bases: (10,2,5,16 etc.)
Division algorithm
Besides studying softwarg for specific mathemat-
Sieve of Eratosthenes
ical areas like the ones just discussed, it is important
Optimizing terms
to consider software which supports specific mathe-
matical methods which have importance in different Summation of arithmetical series according to
areas. Here algorithms in their original sense (think Gauss
of Euclids algorithm) are most familiar and were Fundamental operations with sets
integrated into mathematics education even before
6 Calculation with fractions (handling formal
the advent of computer systems. First we shall give
an example of an algorithmic strand, which fits the rules)
curriculum for the German Gymnasium. From Greatest common divisor and least common
this you will be able to see how this old mathemat- multiple (algorithm of Euclid in several varia-
ical idea of algorithm can be extended to a num- tions)
ber of complex mathematical problems. Then we Prime numbers, twins of prime numbers, distri-
shall discuss the general problem of how to com- bution of prime numbers etc., factorisation of
bine in class teaching students to understand and numbers
execute mathematical methods and to solve math- Arithmetic means, relative frequencies
ematical problems (from multiplication to integra- Diagrams of descriptive statistics
tion) by hand or brain with the need to tell them
that there are computers which can do these things 7 Tables of proportions
easily if you just give the problem to them in the Calculation of percents and interest
proper way. This is the black box/white box prob- Random experiments
lem. Finally, we discuss two more general methods Constructive geometry in two dimensions
of growing importance in mathematics education (as Geometrical mapping
well as in mathematics research) - simulation and
model building. 8 Algorithm of Heron
Iterations for linear equations
The algorithmic strand. Symbolic processing with equations
Algorithms are patterns with a certain schematic 9 Solution of quadratic equations
background; although high mathematical invention
Graphs of quadratic functions
was necessary for their discovery, only stupid and
Combinatorics
exact processing is needed for their application.
With this didactic philosophy the teaching of con- Continuation of geometry (similarity)
cepts and theories of mathematics had priorit#y at 10 Several methods of integration of the circle
schools. The use of algorithms formed the center of
Division of polynomials
exercises, homework and control of achievement and
Trigonometric construction
so pupils were educated as if they were little com-
puters. Related to this secondary role of algorithms Descriptive statistics
is the fact that several thousand years of history of
11 Experiments with sequences and series
mathematics have not produced a uniform language
Discussions of functions
for the description of algorithms. Now there is a
Algorithm of Newton with variations
For a more detailed analysis and critical descrip- Regula falsi
tion see Biehler and Winkelmann (1988). Methods of optimisation
12 Methods of integration according to Simpson, tangent of a function instead of the function itself.
Gauss, Romberg etc. Or you could take tangents with the same constant
Symbolic integration slope as the first tangent (the method converges in
Algorithm of Gauss for systems of linear equa- many cases). Alterations of this kind are generally
tions with variations impossible with acquired programs, which seldom
Operations with matrices etc. allow such open didactical processing. Naturally,
for these purposes the teacher needs a simple and
13 Stochastic simulations transparent computer language with natural key-
Symbolic handling of limits (lHospita1) words and sufficient mathematical operators as well
Standard methods of inductive statistics as a compiler which can understand the language
Methods for numerical, graphical and symbolic in the same sense as humans. Teachers need as
solution of simple ordinary differential equations well a good cooperation with teachers and pupils
of computer science, who can construct good pro-
This listing refers to a basic level of higher ed- grams according to their desires. Some programs in
ucation. For the advanced level (Leistungskurs the school market need to have a didactical dimen-
with 6 lessons in the week) a lot of possibilities can sion so that, for example, the plotting of functions
be added in the last three years such as: can be stopped and continued using the intuition of
complex numbers, special numerical methods, the pupils. During algorthmic processing intermedi-
algorithms of the theory of graphs, fitting of ate suppositions about the results should be possible
curves according to Taylor and Gauss, interpo- which can be verified or falsified.
lation of functions according to Lagrange and
Newton, cubic splines, study of nonlinear iter- Symbolic Processing/Symbolic
ations, mapping and representation in three di- Manipulation
mensions, constructive non-euclidean geometry In recent years symbolic processing for personal
etc. computers has entered into schools (see the chap-
Also in the content of an algorithmic strand, the ter by Hodgson and Muller). Solving linear and
methodological aspects need not be lost. Several quadratic equations, equations of third and fourth
basic formulations of computer algorithms are help- degree, large systems of linear equations, simplifica-
ful for mathematical comprehension, too. For ex- tion of rational expressions with towers of double
ample, pupils always have difficulties understanding fractions, division and simplification of polynomials
the usual notations of sums, double sums, etc. The can all be done with symbolic algebra, often inte-
algorithmic notation using for-loops or nested loops grated with the direct processing of very large in-
removes many difficulties in understanding the role tegers. Where exact methods fail, approximations
of summation index etc. The practice of program- are possible. Symbolic differentiation and integra-
ming recursions is helpful for understanding the log- tion, symbolic vector analysis and, finally, the sym-
ical basis of induction proofs, etc. bolic solution of ordinary differential equations of
For nearly all of the subjects listed software is first and second order together allow the possibil-
available, some of it with more options than are ity of ignoring all the rules of school mathematics
needed in schools. Most teachers are pleased that in a traditional sense. These packages are made by
they do not have to enter into the specialities of professionals. Therefore, they often do not present
graphic representation and the other higher work intermediate steps and some other didactical re-
of computer insiders. Still some of them remember main. Some of the symbolic packages are not pro-
another kind of work only a few years ago: For im- grammable by the user. Nevertheless the union of
portant algorithms of mathematics (Euclid, Gauss, numerical, graphical and symbolical tools has enor-
Newton, Simpson etc.), teachers themselves had to mous power for schools.
write their own programs. The advantage of this
Enlightenment through Black Boxes
was that they could develop the central ideas simul-
taneously in their classes and in the programs. The In a recent article, Buchberger (1990) asks,
disadvantage was that the handling of many pro- Should students learn integration rules?, given
grams was not easy. Still, a further advantage was that now there are computer algebra software sys-
that the teacher could modify an algorithm using tems available which solve any integration problem
the (sometimes unusual) suggestions of the pupils. much more quickly and more reliably than any stu-
As an example, in Newtons method for the solu- dent could ever do with paper and pencil. Buch-
tion of transcendental equations, you could take the berger immediately generalizes the question for all
those areas of mathematics which are trivialized The graph of a function, as displayed by a function
by modern software, especially computer algebra plotter, is different from the graph of the function
systems. He answers - for mathematics and com- as normally defined within mathematics. But the
puter science majors - with his White Box / Black experience of doing function plotting and a reflec-
Box - Principle: Students should learn the theories tion on the possible pitfalls (e.g. vertical asymp-
and algorithms of such an area first, using the soft- totes, discontinuities or the proper determination
ware only for subordinate tasks (e.g. partial fraction of maxima) may help in understanding results and
decomposition) but, after having st#udied the area, becoming aware of possible limitations. For the nor-
all calculations from this area should be left to the mal student it is not worthwhile to learn the special
computer. tricks and algorithms programmers of function plot-
For schools and general mathematical software ters use to give reasonable results even in difficult
the situation is more complicated: Numerical and situations. Analogous knowledge is needed in gen-
graphical oriented software doesnt trivialize an area eral in the use of numerical software - possible pit-
of mathematics, but may provide profound help in falls, trade-offs between step widths and obtainable
solving problems; school mathematics does not only accuracies, between reliability and speed, etc.
provide mathematical theories and algorithms but Algorithmic. Here the user knows - on a cer-
also their intended applications, their modes of use tain level - the specific algorithmic approach used
and the translation schemes needed in using them by the software, for example, that the numerical in-
outside of pure mathematics. High school students tegration software uses Simpsons rule, which the
are just not future mathematicians, but could be use had applied in some hand calculations. But for
regarded as future users of mathematics as well, a suitable use of the software, the user has to have
who obviously should have a different attitude to- some more general knowledge, too - the approxi-
wards mathematical tools. So here we do not give a mation character and the order of the algorithm, its
recipe but rather some considerations which might domain of validity, in what circumstances to switch
help in coping at school level with the problem of to other algorithms, etc.
using ready made software which cannot be made All three kinds of knowledge have their special
translucent, since the details may be too com- value, and in most circumstances they should com-
plex, or totally hidden from the user, or just not plement each other. There is no a priori best way of
worthwhile studying for secondary school students. enlightening a software black box. Of course, math-
First of all, using ready-made mathematics, even ematics teaching has the duty to enlighten black
if not fully understood, is to be seen as taking part in boxes, to make them grey at least, but in which
mathematics as a social enterprise. It may be looked way and to what extent has to be decided in view of
on as part of teamwork: Users rely on professional the intended use of the software, the kind of knowl-
mathematicians and programmers. But the coop- edge to which this new knowledge is to be added
eration is anonymous since the user cant t,alk to and connected, and to the overall goals of mathe-
coworkers, and users have to know a lot in order matics teaching in the specific age group and school
to use the black box correctly and with beneficial system in particular,
results. But knowledge about black boxes (proce-
dures, algorithms, etc.) can be of various kinds: On the Concept and Importance of
Logical or ezternal. The user knows the math- Simulations
ematical specification of the result the software de- How does one simulate a dynamical process?
livers, but doesnt know the method by which it has Such a process is described by specifying the transi-
been achieved. This is the classical black box and tion from one state of the system to the next state;
is usually the case with the use of computer algebra mathematically this is done by (systems of) differ-
systems or simple calculators. A symbolic integra- ence or differential equations. In order to simulate
tion can be understood (and independently checked) such a process, one first has to specify all param-
even if the internal Risch algorithm isnt understood eters, initial states and possible external influences
or its existence even known. The cosine of a number numerically, and then follow the evolution of the
can be interpreted correctly as the best approxima- states numerically, replacing all mathematical oper-
tion within the domain of machine numbers to the ations which have no direct arithmetical translation
correct real number, etc. by numerical approximations. Some ending condi-
Analogous. If a complete specification of the re- tions have to be efficiently specified, too, for exam-
sult of the software is not available, an analogous ple, the maximum number of states to calculate, in
knowledge of a similar algorithm may often help. order to prevent never ending calculations.
74 InfIuence of Computers and Informatics on Mathematics and Its Teaching
The resulting numbers - normally quite a lot - several different initial conditions, solves the result-
can be given as tables or graphics. If the concrete ing initial value problems by numerical methods and
choice of parameter values or initial states is not plots the results. The emerging picture should give
dictated by the situation but just ad hoc in order some insights into the flow-lines of the differential
to be able to start the simulation run, the whole equation, its overall behaviour and possible loca-
process will have to be repeated with other values tions of critical points.
fixed - that means defining another scenario - to Simulations normally share a double experimen-
get an overview over the behaviour of the system in tal character: First by the numerical approxima-
a range of scenarios. tions whose errors can be only estimated since the
From this description we have a geberal infor- assumptions of strict error control in most cases can-
mal definition: By simulation (in mathematics) we not be verified by numerical methods alone, and sec-
understand the effective operational translation of ond by the fixing of the parameters, boundaries, etc.
mathematical objects or processes into numerical Simulations need to be complemented by some the-
operations. (Outside mathematics the concept has ory, however rudimentary, in order to lead to insight
to be extended to include the building of a mathe- and understanding. Thus the plotting of the sine-
matical model first .) function can only give a non-misleading intuition,
Simulation in this sense is a general mathemat- if the continuity and periodicity are known or can
ical method which has always been used but has be abstracted by the consideration of a well chosen
gained importance enormously through the avail- sequence of (simulated) pictures with some zoom-
abilty of effective numerical machines, especially ing or similar means. The insight does not come
computers. As a method it is very often not dif- from the pictures. The intellect of the students has
ficult to apply, and it can be a mighty instrument, to see the connections between the pictures and the
especially if combined with other, more traditional necessities behind them; but to see the facts given
mathematical methods such as proof, construction, by the simulation may strongly help the student to
algebraic calculation, analysis, etc. understand the facts given by some theory.
Here are some examples of simulations:
Model Building
l Function plotting. The mathematical object is
the graph of the function, say of f(z) = sin r, which The building of mathematical models is seen
is a subset of R2. For the simulation one has first by many people as the heart of application ori-
to fix boundaries, say from --7~ to 2z, then approxi- ented mathematics teaching. If done properly, the
mate the interval [-a, 2?r] by a finite set of floating usual restriction to linearity assumptions will soon
point numbers, calculate approximations to the sine be noticed as inappropriate, and the use of simula-
of these numbers, determine screen pixels to corre- tion software in order to explore the (mathematical)
spond to the calculated values, connect those pixels models developed becomes necessary.
by the built-in line-drawing routines, and display Here we describe briefly dynamic model build-
the result. The fixing of parameters will become ing of simple growth processes in the mathematics
even more apparent, if you simulate functions with classroom with the program Modus, which at the
parameters, say f(z) = sin(az), a E R. moment is being tested in schools in a preliminary
l Stochastic simulation. The mathematical ob- version. As with most dynamic modelling tools, the
ject is, for example, a stochastic variable with its crucial concepts are the distinction of the main vari-
distribution, mean and variance, say a uniformly ables as levels and flows. Levels can only be changed
distributed variable transformed by some compli- through flows; this property is described in formal
cated process or function f. To simulate it, you take mathematical language by use of difference or differ-
a finite number of uniformly distributed (pseudo ential equations, the flows being the derivatives of
)random numbers, transform them by (a numerical the levels. The model building is done by construct-
approximation to) f, take the resulting finite dis- ing structure diagrams, thus avoiding the necessity
tribution as an approximation to the distribution for an abstract formal language. The students eas-
sought, and calculate its mean and variance. ily develop linear and exponential models of growth.
l Solution to a differential equation.2 The math- The step from linear to exponential growth is made
ematical object is the general solut,ion to the given by changing the constant flow to a (linear) function
differential equation. To simulate it, one chooses depending on the level, thereby introducing a first
feedback loop (see Fig. 12 and 13).
2 An indefinite integration is a special case of this,
namely the solution of y = f(z).
las for the logistic function, the growth behaviour

can be completely understood from the model itself,
and becomes evident by observing parts of the phase
diagram being generated dynamically (Fig. 15).
Figure 12
turtle there are two more turtles, an axial symmet- than from the final picture. You see that a straight
ric turtle and a central symmetry turtle. They line remains a straight line, you see how the direc-
are controlled in the same way as the standard run- tion changes under axial symmetry and how it re-
ning turtle. But they do not only draw the figure mains the same under central symmetry. You also
controlled but also the figures symmetric image in see that a straight line and its picture are parallel
a different colour relating to the I- or y-axis or the under central symmetry, but have different direc-
centre point of the screen. This happens simultane- tions, etc. Figure 2 contains some examples. Unfor-
ously and pointwise. This mode of construction for tunately, the dynamic quality of the turtles can-
symmetries helps the user to recognise the proper- not be seen from these figures.
ties of the mappings immediately and more easily Figure 3 shows how the following question can
be examined: What happens when reflecting a tri-
angle in different positions relative to the axis of
symmetry or a point?
Figure 4 gives a systematic answer to the ques-
tion, How can quadrilaterals be generated by re-
flecting triangles?
,=
\- /
L
cl3
i
Figure 2 Figure 4
First, it is convenient to choose a side of a trian-
gle as an axis of symmetry. Then with the turtle you
get a kite. The special case of an isosceles triangle
occurs if the angle adjacent to the axis is 90 degrees.
If this angle is greater than 90 degrees, then you get
a quadrilateral which is not convex. You can also
get a rhombus and square by starting with special
triangles. But you never get a general rectangle or
a parallelogram or a trapezoid. The central sym-
metry turtle, however, applied on the centre of a
76 Inf7uence of Computers and lnformatics on Mathematics and Its Teaching
25 pupils in the class. Therefore, often theoretical

questions related to algorithms (additional special
cases, restrictions, possibilities for application etc.)
have to substitute for the real use of the algorithms
for problemsolving in examination periods.
The consequences for the curriculum are very
important. School mathematics was determined for
centuries by the number of accessible methods for
.solving problems - equations of no higher degree
than 2, systems of equations with 3x3 or 4x4 ma-
trices etc. Application problems were selected care-
fully so that powerful computational tools were not
needed. With the speed and the capacity of mem-
ory of modern computers in schools (newer stan-
dard: 80386 processors, 1.2 MB RAM), numeri-
cal and graphical approximations for solving equa-
tions of higher degree and handling matrices of 10
or 20 columns and rows are no problem. Graphi-
cal representation of large sets of higher functions
or of complicated geometrical situations are also no
problem as are the symbolic transformation of com-
plicated rational terms or the symbolic solution of
differential equations with interesting initial condi-
tions. With these tools teachers can leave the small
garden of traditional school problems and amplify
enormously the orientation to modern application.
Let us demonstrate this with two examples.
First, from pure mathematics: After teaching curve
fitting by Taylor approximations or Fourier approx-
imations in the classical manner with the usual
demonstrations you can continue with Pad& approx-
imations using rational functions and use these for
good approximations to functions with singularities
(see Fig. 16).
Our second example is from applied mathemat-
ics: The teacher can show how to compute ap-
proximations to curves of highways in the students
neighbourhood by parametric splines with the help Figure 16
of the computers.
Thus various new fields are opened for the cur- Make students familiar with the basic structure and
riculum. Simulations in natural science and social function of a computer system and teach them how
science using systems of difference equations can be to manipulate it. This situation has changed rapidly
used to solve interesting environmental or economic and will have changed totally in the near future since
problems never before accessible in schools. The most students now get acquainted with computers
theory of graphs or the theory of functions with in their daily lives, in their family and recreational
complex variables are other examples of new ele- environments, perhaps in computer science educa-
mentary work with modern tools. tion, and so on. This means the computer has a
Speculation on the Future new importance in math education, a more fruit-
ful one, more oriented towards mathematics. This
During the first twenty years of computer use is described as follows in a study of mathematics
in schools, the mathematics classroom was the first education for the information age to be realized
place where most students met a computer at all. in the Japanese New Mathematics Curriculum [Fu-
So math teachers had to pursue an additional goal: jita/Terada 19911. In upper secondary schools pri-
ority should be
a nnite numoer of unltormlv dl.d,rlhl:tPd fnppllrlh

on giving to students opportunities to do teacher-initiated use, the student-initiated use and

mathematics rather than improving their the system-initiated use.
techniques. Students should understand that From the viewpoint of a computer-supported
computers are powerful tools for intellectual curriculum, teaching with computers in a classroom
activities by human beings. In this connec- will consist of the following six components:
tion, studying mathematics may be the first 1) trial, where learners are invited to the new
as well as the best experience for students to topic with fun applications offered by the com-
use computers for properly intellectual pur- puter.
poses, namely, to study academic subjects
2) approach, where learners have heuristic and
with computers. These experiences could
operational experiences with the aid of comput-
even be regarded as a prototype of scientific
ers.
research activities with computers. Some
good students will have chances to observe, 3) teaching, where the teachers give a lecture
to model and to analyze in a mathemati- and learners get supplementary review and as-
cal manner various phenomena presented by sistance from computers.
computers. Furthermore, computer simula- 4) experimental understanding, where learners
tion is close to mathematical reality. On the grasp concepts and facts through inductive and
other hand, computers are extremely help- experimental recognition with the aid of comput-
ful in fostering students mathematical liter- ers without being burdened by too much drill.
acy. Rich mathematical experiences offered 5) exercise, where learners can perform adequate
by computers, particularly those through op- exercises at their level and using standard (but
erational work by students, will pave the way interactive) CAI.
for the majority of students to grasp con-
6) survey, where learners review the topic which
cepts and to understand fundamental facts they have learned and are given chances to view
in mathematics. further developments and applications.
The New Curriculum plans three courses, Math- The principal underlying purpose of the New
ematics I - III, in grades 11 - 13 with a total of 10 Japanese Curriculum is to cultivate mathematical
units (1 unit requires 35 class hours of 50 minutes intelligence by aiming at two targets: M&hemat-
each), covering a core of mathematics to be learned ical Literacy and Mathematical Thinking. The as-
by all students with Math III certainly to be learned pects from the curriculum mentioned above show
by all science and technology students. Three more that computer systems are considered to be very
courses, Mathematics A, B, C, in grades 11 - 13, to- helpful for both fields.
talling 6 units, are composed of four option modules These two fields are also mentioned among the
from which two modules are freely chosen for in- principles for the development of a new mathematics
struction by teachers or schools. Module 4 of Math curriculum in the USA by 2000 [Ralston 19901. In
A, computation and computers, offers students the this reference it is stated that Mathematical educa-
chance to get to know and become familiar with tion should focus on the development of mathemat-
computers as a tool for mathematics. Module 4 of ical power not mathematical skills. As to informa-
Math B, algorithms and computers, deals with the tion technology there is this principle: Calculators
powerful function of computers in doing algorith- and Computers should be used throughout the K-12
mic computations in mathematics. Math C is char- mathematics curriculum; moreover, new curricula
acterized by the key phrases application minded and new curriculum materials should be designed in
and do math with computers in the areas of ma- the expectation of continuous change resulting from
trices and linear computation, various curves, con- further scientific and technological developments.
its and polar coordinates, numerical computation Goals from these principles follow for the elemen-
and statistical processing. The study mentions that tary grades (l-6) as well as for the secondary grades
the newly introduced topics related to computers in (7-12). So the teaching of arithmetic in elemen-
Japanese high school mathematics require certain tary schools should be characterized by : . . . a use
preparation for success, namely, purposeful text- of computer software in the teaching and learning
books, effective teacher training, quality software process, proper and efficient use of calculators
and relevant development of teaching materials and for most multi-digit calculations as well as calcula-
methods. Indeed, the educational use of comput- tions involving negative numbers, fractions and dec-
ers in class is non-routine and should be explored imals. One important example of computer use
with respective emphasis of its t,hree aspects; the in the secondary curriculum follows from the goal
that this curriculum should develop students sym- AUTOCALC, Shell Centre for Mathematical Edu-
bol sense. This means developing the ability to cation, University of Nottingham NG7 2RD,
represent problems in symbolic form and to use and UK.
interpret these symbolic representations, and the
CABRI GEOMETRY, Baulac, Y., Bellemain, F.,
ability to identify the symbol manipulations neces-
Laborde, J.M., Laboratoire L.S.D. IMAG
sary to solve problems expressed in symbolic form
Tour Irma BP 53 X, 38041 Grenoble Cedex,
and to carry out manipulations using mental com-
1988ff. Macintosh, IBM PC. Versions in other
putation, pencil and paper, a symbolic or graphic
languages such as English or German are avail-
calculator or a computer.
able.
It was noted above that new mathematics cur-
ricula should be designed in the expectation of fur- DEVELOPING TRAY, Inner London Education
ther technical and scientific developments. Most Authority, NCET, 3 Devonshire Street, Lon-
certainly these will occur in artificial intelligence don.
and telecommunications. In a survey on Technol- EARLY SCIENCE in Exploring Primary Science
ogy and Mathematics Education, James Fey [1990] and Technology with Microcomputers, edited
writes about artificial intelligence, expert systems by Jan Stewart, Council for Educational Tech-
and tutors: One of the very active areas of infor- nology, NCET, 3 Devonshire Street, London.
matics research is exploring ways that computers
GEOMETRIC SUPPOSER, Schwartz, J., Yerushal-
can be programmed to exhibit behaviour that sim-
my, M., et al, Sunburst Communications 1985
ulates human information processing. There are a
ff. Apple II, Macintosh, IBM PC.
number of projects in mathemat,ics education that
are attempting to capitalize on this computer capa- GRAPHIX, Tall, D., van Blokland, P., Kok, D.,
bility to design programs that act, in various ways, Duisburg: CoMet Verlag fiir Unterrichtssoft-
like teachers. The most interesting work along these ware 1989. ISBN 3-89418-862-6. IBM-PC. An
lines is producing intelligent tutors for an array of English version of this program is availablefor
mathematical topic areas including arithmetic, al- example, through Sunburst under the title A
gebra, geometry and proof, and calculus. There Graphic Approach to the Calculus.
are some preliminary indications that those tutors KALEIDOSCOPE, Graf, K.-D., Information can be
provide very effective adjuncts to regular teacher- obtained from the author at Freie Universitiit
directed instruction. Berlin, D-W-1000 Berlin 33, Germany.
As for telecommunications, one might think that
MODUS, Walser, W., Wedekind, J., Duisburg:
this will be important for general or social education
CoMet Verlag fur Unterrichtssoftware 1992.
only. It is likely, however, that the ability to commu-
IBM-PC.
nicate about mathematical problems in a worldwide
group of peers will develop new attitudes towards SEEK, Longman, UK; also Shell Centre for Mathe-
problem solving, different from the widespread sin- matical Education, University of Nottingham
gle attack of scientists and students. Also, it can NG7 2RD, UK.
be imagined that a feeling for the benefits of inter-
SYMMETRIC TURTLES, Graf, K.-D. (see KALEI-
national and intercultural understanding can grow
DOSCOPE)
more intense through cooperation in a serious
field like mathematics or science, in addition to the TIMES in Teaching with a micro: Math 3, Phillips,
effects of leisure fields like music, movies, etc. R. et al, 1986, Nottingham: Shell Centre for
We want to conclude this article by pointing to Mathematical Education. BBC Micro.
one of the greatest problems in the changing of the
Books and Papers
mathematics curriculum under the challenge of com-
puter systems: We must convince the curriculum Abelson, H. and di Sessa, A. [1985]: Turtle Geome-
makers and those who put changes into effect about try: Computation as a Medium for Exploring
the necessity and the advantages of this change. We Mathematics, Cambridge: MIT Press.
hope that this article will provide good arguments Biehler, R. and Winkelmann, B. [1988]: Mathe-
to everybody who wants to tackle this problem. mat&he Unterrichtssoftware: Beurteilungsdi-
mensionen und Beispiele in Schmidt, Giinter
References (Ed.): C om p u t er im Mathematikunterricht,
Der Mathematikunterricht 34, Heft 4, 19-42.
Software ISBN 3-617-24022-4.
Buchberger, B. [1990]: Should Students Learn In- Computer in der Schule 3, Stuttgart: B. G.
tegration Rules?, ACM SIGSAM Bulletin, 24, Teubner, 45-72.
1, 10-17. Winkelmann, B. [1987]: Information Technology
Dubinsky, E. and Fraser, R. [1990]: Computers and Across the Curriculum, in Johnson, D.C. and
the Teaching of Mathematics, selected papers Lovis, F. (Eds), 89-94.
from ICME-6, Nottingham, UK: The Shell Winkelmann, B. [1992]: Themenheft Wachstum.
Centre for Mathematical Education. Dynamische Systeme im Mathematikunter-
Fey, J.T. [1989]: Technology and Education: A richt, Soester Verlagskontor, Soest.
Survey of Recent Developments and Important
Problems. Educational Studies in Mathemat-
ics, 20, 3, 237-272.
Fey, J.T. [1990]: Technology and Mathematics Edu-
cation in Dubinsky and Fraser, 73-79.
Fujita, H. and Terada, F. [1991]: A Coherent Study
of Mathematics Education for the Information
Age; a Realization in the Japanese New Math-
ematics Curriculum, plenary Lecture at ICMI-
China Regional Conference on Mathematics
Education in 1991, Beijing, China. (Prints
available from the first author, Department of
Mathematics, Meiji University, Japan).
Goldenberg, E.P. [1988]: Mathematics, Metaphors,
and Human Factors: Mathematical, Techni-
cal, and Pedagogical Challenges in the Edu-
cational Use of Graphical Representation of
Functions, Journal of Mathematical Behavior,
7, 2, 135-173.
Graf, K.-D. [1988]: Using Software Tools as Ad-
ditional Tools in Geometry Education with
Ruler and Compasses, Education and Comput-
ing, 4, 3, 171-178.
Graf, K.-D. and Hodgson, B. [1990]: Popularizing
Geometrical Concepts: the Case of the Kalei-
doscope, For the Learning of Mathematics, 10,
3, 42-50.
Johnson, D.C. and Lovis, F. (Eds) [1987]: Infor-
Amsterdam: North-Holland.
Okamori, H. [1989]: Mathematics Education and
Personal Computers, Tokyo: Daiichi-Hoki
Shuppan.
Ralston, A. [1990]: A Framework for the School
Mathematics Curriculum in 2000 in Dubinsky
and Fraser [1990], 157-167.
Schumann, H. [1990]: Neue Moglichkeiten des Ge-
ometrielernens in der Planimetrie durch inter-
aktives Konstruieren, in Graf, K.-D. (Hrsg.):
.- ..-....__.m ----- ,--- _-__ _-_ .

A FUNDAMENTAL COURSE IN HIGHER MATHEMATICS INCORPORATING
DISCRETE AND CONTINUOUS THEMES
S. B. Seidman
Auburn University, Auburn, AL 36849, USA
M. D. Rice
Wesleyan University, Middletown, CT 06857, USA
THE CURRENT MATHEMATICS

CURRICULUM: TRADITIONS AND last decade (Ralston 1981, 1989, Kenney and Hirsch,
CONCERNS 1991). Just as a major motivation for the predom-
For many years, a crucial place in the mathemat- inance of calculus in the curriculum has been the
ics curriculum of the last year of secondary school or wide range of applications of continuous mathemat-
the first year of university studies has been occupied ics, the challenge to that predominance has arisen
by the differential and integral calculus. The calcu- from the steadily increasing interest in the applica-
lus can be seen both as the culmination of the sections of discr& mathematics in many disciplines.
ondary school mathematics curriculum and as the This increasing interest in discrete mathemati-
beginning of serious study of mathematics at the cal applications can be primarily attributed to the
university. In some sense, the study of calculus has widespread use of computers. Computers are essen-
become synonymous with the serious study of math- tially discrete machines, and the mathematics that
ematics. The central and essential position occupied is needed to use them is also discrete. As a conse-
by calculus can be traced to at least two interrelated quence, the discipline of computer science is heavily
causes. dependent on a wide variety of discrete mathemat-
For mathematicians, calculus represents the ical ideas and techniques. Furthermore, the easy
methodology and techniques needed for the study availability of computers has encouraged the use
of functions, first defined on the real line, then on and development of discrete mathematical models in
higher-dimensional Euclidean spaces, and finally on many disciplines. For example, operations research
the complex plane. Thus, the study of the calculus models (linear programming, integer programming,
allows students for the first time to acquire the for- etc.) are widely used and are based on a discrete
mal abstract tools that are essential for the further mathematical perspective.
study of much of higher mathematics. It is natural to expect that the rapid growth
On the other hand, calculus provides the foun- of interest in discrete mathematics and its appli-
dation for many applications of mathematics to the cations, fueled by the explosive developments asso-
physical sciences and engineering. These applica- ciated with computers, should have an impact on
tions date back to Newtons original development t,he mathematics curriculum. Although this im-
of the calculus in the seventeenth century, and since pact would have been significant under any cir-
that time they have been wildly successful across a cumstances, its effect in the United States has
vast collection of disciplines, even including (in re- been magnified by other questions that have been
cent years), the biological sciences and economics. raised in recent years about the teaching of calcu-
All of the calculus-based applications are based on lus. Widespread dissatisfaction has been reported
mathematical models that can be regarded as being with the nature of the calculus courses and the
continuous; that is, the quantities being nlodeled knowledge of the students that have completed them
are real numbers (or elements of some Euclidean (Lochhead 1983, Steen 1983, Douglas 1986, Steen
space Ii?). 1988). The computer is also directly influencing the
Given both the central mathematical positzion of Content of the calculus course itself, both by en-
the calculus and its vital role in applications (not couraging the inclusion of numerical methods and
to speak of the interaction between these two fea- by suggesting that symbolic manipulation software
tures), it is easy to see why the calculus has occu- may make emphasis on techniques of differentiation
pied such a fundamental and unassailable posit#ion and integration obsolete (Bushaw 1983, Wilf 1983,
in mathematics curricula. During t,he past several Nievergelt 1987).
decades, however, the central role of calculus has In summary, both the nat.ure of the calculus
been seriously questioned, and the questions have course and the fundamental position that calcu-
been repeat.ed with particular emphasis during the lus has occupied in the mathematics curriculum for
80
Discrete and Continuous Themes 81
more than a century have come under serious chal- those students concentrating in computer science.
lenge. These challenges have come both from within Ralstons proposal has led to substantial discussion
and outside the community of mathematicians, and in the United States on the proper place of discrete
they can primarily be attributed to the increasingly mathematics in the curriculum (Ralston and Young
broad role that computers are playing in the various 1983). The debate has focused on whether discrete
scholarly disciplines represented within the univer- mathematics should precede or follow the calculus
sity and in the wider world. In the next section of in the curriculum of the first two years. Many of
this paper, we will look at the responses that have the arguments advanced on either side are adminis-
been proposed to these challenges. trative in nature, dealing either with the demands
of other curricula (such as physics or engineering)
RESPONSES TO THE CHALLENGE OF or with articulation with other institutions (such
DISCRETE MATHEMATICS as high schools, junior colleges or universities that
have retained the standard curriculum). One result
When any curriculum is confronted by a. new
of this debate has been the publication since 1985
topic that should be included, there are essentially
of over 40 discrete mathematics texts for freshman
two potential responses. The new t,opic can either
or sophomore courses (e.g. Ross and Wright, 1988,
be encapsulated in a course that is added to the cur-
Maurer and Ralston, 1991).
riculum, or it can be incorporated as a fundamental
constituent of a revised course. Most topics that Whether calculus is placed before or after dis-
have been added to the mathematics curriculum in crete mathematics, it is by no means clear that
recent decades have been added as new courses (e.g. students who have completed both courses will be
abstract algebra and topology). able to combine their discrete and continuous math-
It was therefore natural that when mathematics ematical skills in an effective manner. This problem
faculties were asked to include discrete mathemat- has been recognized by some designers of proposed
ics in the curriculum, this was most commonly done curricula, and consequently their calculus propos-
by developing new courses in discrete mathematics. als generally include some discrete aspects, such as
extended discussion of numerical methods and sub-
Such courses were designed primarily for students of
computer science. There have been two fundamen- stantial use of sequences (see, for example, Bushaw
tal problems with this approach. First, the discrete 1983).
mathematics courses were too often taken by third- Another possibility, which has been given little
year students, so that the material was learned too serious attention, would be to develop a new, uni-
late to be of use in the data structures courses taken fied curriculum that would interweave discrete and
by first and second year students of computer sci- continuous themes throughout its courses. While
ence. Second, when students were expected to use the first year of the curriculum would correspond
their discrete and continuous mathematical skills to the calculus course, its real thrust would be the
in fourth-year computer science courses (for exam- study of functional behavior and functional repre-
ple, in the analysis of algorithms), most have found sentation. The course would consider discrete func-
it very difficult to combine these skills effectively. tions (sequences) along with continuous functions,
Many students do not see any connections between and would constantly emphasize analogies and par-
discrete and continuous mathematics, and are un- allels between discrete and continuous situations.
able, for example, to apply calculus techniques to Thus the first year of the curriculum would be pri-
estimate growth rates of discrete functions or to es- marily continuous, but with a strong discrete flavor.
timate the size of discrete sums. This inability to The second year of the curriculum would focus on
combine discrete and continuous skills is also found structure, and would be primarily discrete, but with
in students of probability, operations research and a strong continuous flavor.
signal processing. This paper will argue that a curriculum unifying
Both of the above reasons suggest that discrete discrete and continuous themes is not only feasible,
mathematics should be incorporated as a compo- but has the potential of providing students with a
nent of the fundamental mathematics course that broad, powerful perspective embracing the mathe-
is offered to all students in their first two years of matical ideas and techniques that are needed for the
university study. This suggestion was first made study of computer science. This perspective would
by Ralston (1981), who proposed that the study of also yield a strong mathematical foundation for the
discrete mathematics precede the study of calculus. study of engineering, the physical sciences, and in-
He argued that such an organization would benefit deed for the study of higher mathematics itself.
virtually all students of mathematics, and not just Furthermore, the development of such a cur-
82 Influence of Computers and lnformatics on Mathematics and Its Teaching
riculum would force a reexamination of the top- Models demonstrating the need to construct
ics taught in the conventional calculus course. As functions and to perform curve fitting will be
mentioned above, various recommendations have included.
been made to remove or include particular top-
B. Behavior of discrete functions
ics. Although each such recommendation has been
1. Sequences: Iteration and Recursion
cogently argued, no consistent rationale has been
This section will discuss a variety of sequences
given for the collection of topics that together make
up the proposed calculus course. The first-year including geometric squares, the Fibonacci se-
quence and the sequence generated by the Eu-
course outlined below has a consistent theme - func-
clidean algorithm.
tional behavior and representation - and each topic
to be included in (or excluded from) the course 2. Difference Operators
should be judged on the degree that it matches the The difference operation A will be introduced as
courses perspective. a function on sequences. The recursion scheme
In the following section, a detailed outline and Uk+l - tik = Auk
discussion will be given only for the first year of the
proposed two-year curriculum. At the conclusion of
the paper, we shall return to the second year of the will be treated in order to emphasize special
curriculum, as well as to the larger issues raised by functions defined on the integers. Formulas for
the question of articulation with other curricula. higher differences will be discussed.
3. Summation
A FIRST-YEAR CURRICULUM INCOR- The primary topic here will be the binomial the-
PORATING DISCRETE AND CONTINU- orem, both in its standard form and in the ex-
OUS THEMES pression for (1 +A). The second form will allow
various formulas for finite sums to be presented.
The fundamental thrust of the proposed first-
year curriculum is the behavior and representation 4. Order Notation (0, o) and Limits of Sequences
of functions. Roughly, the first semester is devoted
to tools for the description and analysis of functional C. Behavior of continuous functions
behavior, with the focus shifting to representation of 1. Limit Heuristics
functions in the second semester. Before presenting Limits of functions will be discussed only in
a more extended discussion of the benefits to be terms of limits of sequences. The continuity con-
achieved by including both discrete and continuous cept will be introduced. The operator
topics, it will be useful to give an annotated outline
of the first semester curriculum. A,f = f(z + h) - f(x)
h
A. Functions
1. Number and Relations will be introduced. Analogies to the discrete
difference operator discussed above will be pur-
A knowledge of set concepts and notation is as-
sued.
sumed. Inequalities will be emphasized.
2. First Derivative
2. Functions and Operations
The derivative will be defined, and interpreted
The function concept and functional notation
using tangent lines. It will be shown that differ-
will be introduced, stressing the algorithmic in-
entiable functions are continuous.
terpretation of the function symbol f. Discus-
sion will include domain and range, operations 3. Differentiation Rules
on functions (arithmetic operations, composi- Powers and roots; product, quotient rules.
tion, translation), and graphs of functions. Use- 4. Monotone Functions and Local Extrema
ful functions will be introduced [polynomials, A rigorous treatment will be postponed. Curve
rational functions, exponential functions (de- sketching will be introduced here and the use of
fined on the integers), absolute value, floor, ceil- graphing calculators will be stressed.
ing]. 5. Second Derivative
3. Models Concavity will be discussed and applied to
Algorithms and elementary complexity analy- curve sketching again using graphing calcula-
sis will be introduced (including binary search). tors.
This will allow discussion of the function [lg(n)J . 6. Extreme Values
Maximum-minimum problems will be solved. Substitution techniques will be discussed, as well

Examples will also demonstrate the use of piece- as the use of integral tables and symbolic calcu-
wise linear functions. lators.
7. Related Rates 6. Evaluation of Integrals: Numerical Techniques
The chain rule will be presented, and related rate The trapezoidal rule and Simpsons rule will be
problems will be solved. discussed. It will also be shown how integrals
D. Estimation and error can be estimated using inequalities, and how
1. Mean Value Theorem sums can be estimated using integrals.
Monotone functions will be discussed more rig- 7. Applications of Integration: Aggregation
orously, and the MVT will be applied to global The applications to be treated include work and
estimation of functions. volume.
2. Solution of Equations 8. Applications of Integration: Modeling
Newtons method will be discussed from both ge- The primary theme here will be the recognition
ometric and iterative perspectives. An elemen- of Riemann sums in differing situations. Exam-
tary treatment of error estimation will be given, ples will be taken from arc length and fluid flow.
and critical values will be estimated. The basic point will be that when a model gen-
3. Interpolation erates a discrete (Riemann) sum, it can then be
Interpolation of functions by straight lines and approximated by a definite integral.
parabolas will be discussed using the difference Although this annotated outline gives a good
operators developed above. overview of the first semester of the proposed course,
4. Approximation it is too brief to show how the interweaving of dis-
Second-order Taylor polynomials will be used to crete and continuous themes can lead to major bene-
approximate functions, and the estimated error fits. The following examples are meant to be typical
will be computed. Analogies will be drawn be- of the perspective that will be possible within this
tween interpolation and approximation and be- course structure.
tween differences and derivatives. Example 1: At the beginning of the course, the
discrete exponential function, f(n) = 2, will be in-
E. Integration
troduced, along with its one-sided inverse, g(n) =
1. Introduction
max{lc12k < n} = [lg(n)J. The function g(n) is vi-
The summation operator for sequences will be
tally important in computer science; for example,
introduced. Its relation to the difference oper-
g(n) + 1 is the worst-case number of comparisons in
ator will be discussed. It will be treated as an
a binary search of a list of length n. The growth rate
aggregation operator, and used to motivate the
of g(n) is important, and is usually treated (via cal-
discussion of area.
culus) using LHospitals rule. We suggest a discrete
2. The Definite Integral approach, based on the binomial theorem. Clearly
This will first be introduced using a piecewise Zg() < n, so that g(n)/n < g(n)/2g(). To de-
linear definition. This definition will then be termine the behavior of g(n)/2g() as n -+ 00 it
applied to step functions. The area definition is sufficient to consider powers of 2 since g(n) is
will then be presented, and applied to parabolas constant between successive powers of 2. Since for
using the results on finite sums obtained above. 12 = 2k,g(n)/2 g(n) = Jz/~~, it is only necessary to
Some elementary properties of the definite inte- look at the behavior of k/2k as k: + 00. By the bino-
gral will be presented, including the mean value mial theorem 2k = (1 + 1) 2 Ic(B - 1)/2, and hence
theorem for definite integrals. lc/2k < PL/k(k - 1) = 2/k(k - l), which gives the
3. The Indefinite Integral result that g(n)/2g() ---* 0 and, therefore, so does
This will be explicitly computed for step func- s(n>ln. The simplicity of the discrete argument
tions, piecewise linear functions and parabo- should aid the student in learning, understanding
las. an assimilating the growth rate of the continuous
4. The Fundamental Theorem of Calculus logarithm.
This will be derived from the mean value the- Example 2: The syllabus outline has referred to
orem for definite integrals. The chain rule will analogies between the discrete difference and sum-
be applied to investigate some properties of the mation operators on the one .hand, and differen-
integral of l/2. tiation and integration on the other. For exam-
5. Evaluation of Integrals: Analytic Techniques ple, the difference operator is defined on the se-
84 Influence of Computers and lnformatics on Mathematics and Its Teaching
quence {un} by Au,, = u,+r - u,. If we de- study of functional behavior, and both discrete and
fine the falling factorial function on the integers by continuous functions are treated throughout. Each
x(m) = z(t - 1). . .(z - m + 1) then it is easy to see class of functions is used to develop tools and sug-
that AZ(~) = mx(m-l), A2x() = m(m- 1)~(-2), gest analogies that will be useful for the study of
and finally that Amz(m) = m! and Am+lz(m) = 0. functions of the other class.
Thus the behavior of the difference operator (and its The second semester of the course further elab-
iterates) on the polynomials z:(m) is strongly analo- orates the functional perspective. Rather than give
gous to the behavior of the differentiation operator a detailed, annotated outline, we shall discuss the
(and its iterates) on the polynomials {z}. Further- topics to be covered and describe how they relate
more, since each collection of polynomials provides to the themes developed during the first semester.
a basis for the vector space of polynomials of degree The second semester is primarily devoted to mate-
at most n, an example has been introduced which rial taken from two broad categories, special func-
will be useful in a later course in linear algebra. tions and representation of functions.
One further benefit of the use of difference oper- Exponential and logarithmic functions will be
ators is the natural observation that A2 = 2, or treated in depth. The natural logarithm will be
more generally that Akn = (k- 1)k. This suggests introduced using the definite integral, and its prop-
that exponential functions, whether discrete or con- erties will be investigated. The inverse of the log-
tinuous, may have a special role to play with respect arithm will be motivated using growth models and
to difference or derivative operators, and serves to the differential equation dy/dx = ky and the rela-
motivate the later observation that d/dx(ez) = e. tionship of this inverse to the exponential function
Example 3: The first two examples used dis- will be motivated using difference equations and the
crete ideas to motivate continuous concepts that are discrete logarithm. Finally, the properties of the
to be introduced later. In this example, continuous function e will be developed. Numerical estimates
techniques are used to obtain a discrete result. The for exponential and logarithmic functions will be
identity giving the sum of a geometric progression, used throughout the discussion.
n-1 The next major topic will be trigonometric func-
xk - xn - tions. Here the primary motivation will come from
c x-l
k=O the geometry of the circle and from models of cir-
can be differentiated using the quotient rule to ob- cular and harmonic motion, although discrete pe-
tain the identity riodic functions, such as mod n, will also be used.
The properties of the trigonometric functions will be
n-1
kxk= (n-1)x+-nx:+z developed. Integration by parts will be introduced
c (x-1)2 and applied to the special functions. The special
k=l
integrals leading to the inverse trigonometric func-
Using this identity, it is immediate that tions will be introduced here. Mathematical models
n-1 suggesting the use of trigonometric polynomials will
c k2k = (n - 2)2 + 2 also be used.
k=l Once the standard functions have been treated,
and that it will be natural to discuss various forms of in-
n-1
finitary behavior. The discussion will begin with
k2-k = 2 - !%!
c p-l a reconsideration of infinite sequences, including a
k=l
presentation of indeterminate forms and their ap-
The last result yields plications to order notation. The remainder of this
section will be devoted to improper integrals and
k2-k = 2 infinite series, emphasizing the analogies between
2
k=l these two forms of infinite summation.
since k/2 + 0 as k -+ 00 (see Example 1). This At this point, the focus will shift somewhat from
example serves to remind students that continuous functional behavior to functional approximation and
techniques can be important in discrete situations. representation. Thus the next major topic will be
These examples demonstrate that the proposed power series, with particular emphasis on the use
course does not merely insert a collection of im- of Taylor series to represent functions. Generat-
portant discrete topics into the calculus course, but ing functions for simple recursions will be discussed,
rather expresses a consistent approach to all of the and a certain amount of attention will be devoted
subject matter. The fundamental perspective is the to computational issues and the estimation of er-
ror terms. The constant theme will be the use of course designed for those students.
Taylor series as function approximations to obtain Although much vitally important mathematics
information about functional behavior that would can be subsumed under the general heading of
otherwise be difficult to obtain. functions, an equally important heading is that of
The final topic will be trigonometric series, with structure. While the proposed course is intended
particular emphasis on the representation of func- to give students the most important tools that come
tions using Fourier series. The treatment of Fourier under the former heading, it does not address the
series at this early point will require the introduc- latter. For students of computer science, both head-
tion of complex numbers, which will reinforce the ings are equally important, and thus an important
students geometric understanding of trigonometric place in their education must be found for struc-
functions. Furthermore, the availability of Taylor ture. Much of the debate summarized above on the
series will permit an analytic as well as a geomet- place of discrete mathematics in the curriculum can
ric discussion of the identity eiZ = cosx + isin x. be seen as a debate on the place of structure in the
Finally, the early introduction of Fourier series will curriculum. Following on the first-year course that
make it possible to discuss discrete Fourier series has been outlined above, it is reasonable to develop
and their acplications at a far earlier point in the a second-year course focusing on structure.
curriculum than is presently possible. Such a course will not be described in detail here,
Clearly, the focus on functional behavior and but it is possible to discuss briefly what general top-
representation has produced a first-year course that ics might be included. The primary strands might
is quite different from what is currently taught. The be discrete mathematics, linear algebra and proba-
essential core of the current calculus course has been bility theory. Discrete mathematical topics could in-
retained, but it is always made clear that it is there clude relations, graphs, Boolean algebras and formal
because it throws a powerful spotlight on functional languages. The discussion of linear algebra could in-
behavior and representation. clude some multivariate calculus, which could then
Conversely, many traditionally taught topics be applied in the probability portion of the course.
have been removed. This pruning was only possible Just as with the first-year course, the topics in-
because the developers approached each topic with cluded in the second-year course should be chosen
the same question: How does this topic impact on because they illustrate vital structural themes or be-
the main theme of the course? cause they are motivated by or permit the develop-
Now that the course has been outlined, it re- ment of important applications.
mains to show how it will fit into the curriculum. We The introduction of courses designed along these
will also have to pay some attention to the second- lines will not be a simple matter. The obstacles that
year course that will follow this course, and also will be found range from the need for new textual
to the political and institutional problems that its materials to the difficulty of articulating the new
adoption would pose. courses with other courses and institutions on all
levels. It would be an unfortunate mistake, how-
IMPLICATIONS FOR THE CURRIC- ever, to conclude that because of the certainty of
ULUM encountering what seem to be insuperable obstacles
to the introduction of a truly new curriculum, the
The first question to be addressed is the audi-
only possible strategy is one of incremental change.
ence to be served by the proposed course. It is
The development and introduction of a curriculum
clearly ideally suited for students of computer sci-
integrating discrete and continuous ideas is an ex-
ence, since it merges themes from continuous and
citing challenge, and one that should be taken up in
discrete mathematics in a synergistic manner. Stu-
several places. What is really needed is a collection
dents who have successfully completed the course
of design and development experiments, performed
can be expected to handle the mathematics arising
in out-of-the-way protected environments. Once
(for example) in the analysis of algorithms. It can
also be argued that this course would be well suited a new curriculum has proven its viability and worth
as a first course for students of mathematics, the in one or more of these experimental environments,
it will be time to address the structural and institu-
physical sciences and engineering. For these disci-
tional issues involved in transplanting the successful
plines the major omission has been vector geometry
curriculum to less protected situations.
and multivariate calculus. In many universities, a
large proportion of this material is treated in the
second year, and it is not unreasonable to suppose
that even more could be shifted to a third-semester
REFERENCES
Bushaw, D. [1983]: A two-year lower-division math-
ematics sequence in The Future of College
Mathematics ed. A. Ralston and G.S. Young,
pp 111-118. New York: Springer- Verlag.
Douglas, R. (Ed.) [1986]: Toward a Lean and
Lively Calculus, Proceedings of a Confer-
ence/Workshop at Tulane University, Wash-
ington, DC: Mathematical Association of
America.
Kenney, M.J. and Hirsch, C.R. (Eds.) [1991]: Dis-
crete Mathematics Across the Curriculum, K-
12, 1991 Yearbook of the National Council of
Teachers of Mathematics, Reston, VA: CTM.
Lochhead, J. [1983]: Th e mathematical needs of stu-
dents in the physical sciences in The Future of
College Mathematics, eds. A. Ralston and G.
S. Young, pp. 55-69. New York: Springer-
Verlag.
Maurer, S. B. and Ralston, A. [1991]: Discrete Algo-
rithmic Mathematics, Reading, MA: Addison-
Wesley.
lege Education: the HP-28C, Amer. Math.
Monthly, 94, 895-902.
Ralston, A. [1981]: Computer science, mathemat-
ics and the undergraduate curriculum in both,
Amer. Math. Monthly, 88, 472-485.
Ralston, A. and Young, G.S. (Eds.) [1983]: The
Future of College Mathematics, New York:
Springer-Verlag.
Ralston, A. (Ed.) [1989]: Discrete Mathematics in
the First Two Years, MAA Notes No. 15,
Washington, DC: Mathematical Association of
America.
Ross, K. A. and Wright, C.R.B. [1988]: Discrete
Mathematics, 2d Ed, Englewood Cliffs, NJ:
Prentice Hall.
Steen, L.A. [1983]: Developing mathematical ma-
tunty in The Future of College Mathematics,
eds. A. Ralston and G. S. Young, pp. 99-107,
New York: Springer-Verlag.
Steen, L.A. (Ed.) [1988]: Calculus for a New Cen-
tury, MAA Notes No. 8, Washington, DC:
Mathematical Association of America.
Wilf, H.S. [1983]: Symbolic manipulation and algo-
rithms in the curriculum of the first two years
in The Future of College Mathematics eds. A.
Ralston and G. S. Young, pp. 27- 40, New
York: Springer-Verlag.
TEACHER EDUCATION AND TRAINING
Bernard Cornu
Institut Universitaire de Formation de Maitres, Grenoble, France
Introduction
During the ten last years, considerable progress ing activities for their pupils. These teachers some-
has been made in the development of computer how got the training they wished (even if they learnt
hardware and software, and many valuable educa- a lot by themselves!). But we now need to go fur-
tional experiments have been carried on. However, ther, and the way the use of computers was devel-
computers are not so commonly used as one might oped with some teachers is certainly not applicable
have expected. In many schools the computers are to the teachers. We need to imagine new ways such
locked in a special room, and it is not easy for teach- that all teachers will be able to use computers.
ers to use them. They must plan in advance, be sure In most countries computer science is not yet a
the room is available, get the key and check and pre- school subject. Therefore, except in some particu-
pare the computers. Then they go to the computer lar cases, we do not need to train computer science
room with the pupils, and the time spent there is teachers, but we need to train teachers in all sub-
generally not totally integrated into the course. jects in the use of computers and new technologies
Thus even when the computer is used, the im- in the teaching of their subject. Thus we need to
pact on the learning is not clear. For some pupils, reflect on the contents of such training.
it is clearly useful and fruitful, but do we know why The main problem, as noted, is the of general-
and how? We all know very good and enthusias- isation. We know how to train some teachers but
tic teachers using computers, and they generally do we now need to train all teachers. We have done
it with success. But it is time consuming, it needs some very specific and sophisticated training; we
a great personal investment, and the conditions of now need training which can be easily generalised
success are not easy to reproduce or to transfer to and delivered to all. We must take into account
another situation. the willingness and the abilities of the standard
However, computers are now very common in so- teacher, and design adequate training. The usual
ciety; they are used in many domains of daily life. In training for good and enthusiastic teachers is cer-
many countries national plans for computer equip- tainly not directly reusable.
ment in schools have been achieved, and so a lot This is both a pre-service and an in-service mat-
of computers are available in schools. Much educa- ter. In the next ten years in most countries, one
tional software has been produced, and it is often third of the teachers will be changed (because of re-
of high quality. The use of computers does indeed tirements and the increasing numbers of teachers).
become easier. So pre-service training will be efficient for this third.
Five or ten years ago, the focus was on the devel- The other two thirds will need in-service training
opment of hardware and software, and on original during the same ten years.
experiments in using computers in education. Now In the long term, one must think about the
it appears that teacher training is the next major link between pre- and in-service training. In an
and unavoidable step but, one which has not been ever changing world, it is impossible to give future
sufficiently studied. Most countries are now asking teachers the abilities and knowledge they will need
how to train all current and future teachers in the throughout their careers. They will have to learn, to
use of new technologies for education. think and to reflect continuously. Pre-service train-
Of course, training plans have already been ing is not intended to avoid in-service training, but,
tried. The first ones were generally training in com- on the contrary, to prepare for it! Increasingly in-
puter science. Teachers from various subjects were service training should be considered a normal part
trained in computer science, and one thought that of the job of a teacher. It should not be only for
they would then be able to use computers in their volunteers, but for all!
teaching in an efficient way. It did happen but only
in some cases! And it did not solve the pedagogical The evolution of teaching
problems of the use of computers, which increasingly
appear to be essential. For several different reasons, teaching is going
The use of computers in educat,ion has relied to evolve:
mainly on some enthusiastic teachers who spend l Technology is evolving quite quickly. Hardware
nights and weekends writing programs and prepar- is becoming smaller and cheaper and more and
87
easier to use. Software is also evolving, becoming sometimes individually with pupils. Activities with
more user-friendly, and it is becoming possible to the pupils may occur not only in the classroom, but
use computers with good software which requires also in other rooms of the school such as a resource
little preparation. center, laboratory, a room for small groups etc.
l Pedagogy is evolving. One reason is because re- Altogether the teacher has to be a counselor, ad-
search in education provides better knowledge of visor, organizer, leader and a manager. The task is
teaching and learning. And new tools are begin- not only mastering of teaching, but also mastering
ning to be available for teaching. But pedagogy the management of learning.
is also evolving because of the democratisation of The way teachers work every day is evolving.
education. More and more children have access They will probably be in the school all day and all
to education, and so pupils are increasingly di- week long and will use various tools in preparing
verse, and need pedagogy adapted to their needs. lessons and in teaching. They will work together
In short, teaching needs to be more individu- with colleagues, and even teach together with col-
alised. leagues. Their personal work will also evolve and be
Current and future teachers must be prepared more diverse. The evaluation of pupils is going to be
for this evolution. It is not enough to master more and more complex, and the role of the teacher
the knowledge and some pedagogical strategies and in evaluation will be more important. Evaluation
tools. Teachers must be able to deal with all the itself is becoming more precise and more technical;
evolution which will happen, and to adapt to many the use of evaluation in training and in individu-
different kinds of pupils. alization of education will be a major role for the
The school of tomorrow may be quite different. teacher.
It will be organized according to a variety of peda- Teachers will also have to be involved in the elab-
gogical styles. There will be large rooms for large oration of pedagogical tools. The evolution of teach-
audiences; standard classrooms; rooms for group ing needs new tools, but also new ways in designing
work; rooms for individual work; rooms for practi- the tools. Textbooks, software, video and audio doc-
cal work or workshops; resource rooms etc. Not only uments and resources for pupils will all have to be
will the pupils be provided with a variety of rooms, better adapted to specific pupils or groups of pupils.
but so also will the teachers. Teachers now have Their elaboration will need more techniques, more
a rather standard way of working. They come to technology and more professionalism.
school to give their lessons, and they stay at home to Team or group teaching will become more fre-
prepare the lessons or to mark homework. One can quent. Teachers will work and reflect together and
imagine that teachers will increasingly work with this will soon be considered as a normal component
their colleagues and that they will need to have of the job of a teacher. As intellectuals, teachers
special tools and materials available for their work. must continue training and reflection throughout
Certainly offices must be provided for teachers, and their professional life.
rooms for group work. They will also need labora- Thus teaching can no longer be considered only
tories to prepare lessons using technology. as an art; it is a profession with all the components
The school of tomorrow will be equipped with of the professionalism. And this has consequences
advanced technology - computers, multi-media refor the education and training of teachers. We must
sources, easily available in each room (perhaps with train professionals!
permanent equipment, or possibly by plugging in A good professional must have access to the best
portable machines); Resource centers will also be and most efficient tools and must be prepared to
necessary in schools. Libraries with books, soft)ware, use these tools, to choose the tools to be used and
audio, video, and multi-media products. As is the to adapt to new tools. Once again we note that
case already with other subjects , one can imagine continuous training is a natura.1 part of the job of a
that in the near future, mathematics laboratories teacher.
will be available in most schools. Will the computer be able to replace the teacher?
The role of the teacher is also changing. Since The answer seems to be no. Teaching and learn-
pupils are more and more diverse, the teacher has to ing are very complex processes, and, although tech-
intervene in many different ways, not. only as a lec- nology brings new tools, the main didactic actor is
turer, giving lessons and delivering knowledge to the the teacher who manages the learning process and
pupils. In the classroom teachers must use different adapts it to each pupil and who insures the social-
pedagogical styles and different kinds of activities. ization of the knowledge and its compatibility with
They must also work with small groups of pupils and the external world. Among the productions of the
Teacher Education and Training 89
pupils in the class, those must be identified which But what about new technologies and comput-
deserve status of knowledge. Thus the teacher in- ers? They are linked with each of the aforemen-
stitutionalizes knowledge. tioned competencies. One must think about the role
Of course, personal and individual moments are of the computer with respect to the subject itself.
possible and necessary in learning, and the com- What is the influence of the computer on mathe-
puter can make them more efficient As well, some matics, on the way mathematicians work and on
particular topics can be learned automatically the mathematics which is taught in schools? What
with a computer (for example, typing; training in is the place of computers in the way mathematics is
repetitive skills such as computation, learning by used in society? What is the influence of computers
heart, etc.). But teaching and learning need in- on the pedagogy and didactics of mathematics? On
teraction between the teacher and the pupils, and evaluation? What is the role and the use of the com-
among the pupils; the computer must be used in a puter in class management, in individualization, in
way that facilitates this interaction. the organization of the teaching? How does it affect
the psychology of the pupil? What technical help
Teacher training can the computer bring to the teacher?
Of course, there are no definite answers to these
Which competencies?
questions; the education of teachers must make
In talking about teacher training we need first to them able to ask these questions and reflect about
determine the competencies which are necessary for them. Education cannot give definite competencies,
a teacher. Which teachers do we need for tomorrow? but it must give an aptitude to evolve; it must give
Which kinds of teachers do we want, to prepare? the basic tools necessary to be able to build ones
First, we certainly need teachers who have mas- own strategies, ones own answers.
tered perfectly the knowledge they will have to We now try to list some of the competencies a
teach; teachers must be competent in their subject. teacher needs in computers and computer science,
But this is not enough. They must not only be good remembering that our purpose is not to train com-
in their subject; they must also be good about puter science teachers, but mathematics teachers.
their subject. They need to know about the origins l Basic tools: such as word processing, spread-
and evolution of their subject, about its history and sheets, data processing, and also other techno-
epistemology. They need to know about the role of logical tools such as video and the overhead pro-
their subject in society and about its applications. jector. This is certainly a very important point:
They need to know about the philosophy of their If we want ALL teachers use new technologies,
subject. they must be totally familiar with the most com-
We need teachers who are able to communicate mon and easy to use; it is an absolute necessity
knowledge and to make pupils construct their own that teachers be able to use computers for ele-
knowledge. Teachers must be educated in the peda- mentary applications. This is the way to make
gogy and didactics of their discipline. They must the computer part of the daily life for teachers.
know about the obstacles to learning, they must l Technical elements: to be able to use the hard-
know about the errors students may make and their ware, to manipulate the main accessories, to
role in dealing with these errors. They must know identify elementary troubles, and to deal with
about the conditions which facilitate learning and the technology in the school; teachers need a ba-
they must know about evaluation. sic level of familiarity with technique.
We need teachers able to manage and lead their l Elements of computer science. But to what
classes. They must know about groups and individ- extent? Teachers certainly need to know just
uals. They must have some knowledge of psychol- enough in order to understand what happens;
ogy. but the links between mathematics and com-
We need teachers able to advise and orient their puter science are so strong that it is certainly
pupils. Thus they must know the educational sys- useful to know about some fundamental con-
tem, its place and its role in society so they need cepts (as well as some concepts of algorithmics -
some knowledge of sociology. see the chapter by Maurer).
We also need teachers trained in the technical l Mathematics and informatics. Mathematics is
aspects of their job, able to speak loudly and clearly evolving and changing under the influence of
enough and able to use technical tools etc. computers and informatics. Therefore, teachers
In general then, teachers have many different need to maintain their mathematics knowledge
roles and must be competent in each of them and to practice mathematics from an informatics
viewpoint. Mathematics is becoming more ex- So the most important thing in educating teach-
perimental, more algorithmic, more numerical; ers how to use computers in teaching is not to give
teachers must be able to follow the evolution of lectures on how to use computers, but to actually
mathematics, and to acquire new competencies use the computer in the training. This is true for
and new attitudes and to be able to carry out all new technologies. You should use the overhead
new activities in mathematics. projector in the training, rather than give a lecture
l Using existing resources. Teachers must be able on how to use the overhead-projector.
to know what exists - different software, differ- If you want to convince teachers that pupils can
ent tools, different strategies for teaching. They learn better with the computer, just make these
must be aware. of new products which appear. teachers or future teachers actually learn something
They must be able to choose among existing re- with the use of computers.
sources according to the needs of their pupils This means that the training should include ac-
and according to their pedagogical choices. They tive parts, even if some theoretical aspects are also
must be able to advise pupils which products necessary. One often says that problem solving is a
they should use. good way to learn mathematics; similarly, the solv-
l Pedagogy, didactics and the computer. One of ing of teaching or learning problems is a good way
the main problems of the use of computers for to learn about pedagogy, and the solving of teach-
mathematics teaching is the integration of the ing or learning problems using new technologies is
computer activities into the pedagogical strat- certainly a good way to learn about the use of new
egy. Too often, computer activities are just technologies in education.
added to the usual lessons. An optimal use of Teacher training should not be only an accu-
the computer needs not only good knowledge mulation of knowledge. As already noted, teach-
of the hardware and software to be used, but ers should be prepared to evolve and adapt to new
also mastering of the problems of learning. A situations.
teacher should be aware of what we now know Among the different methods which can be used
about how pupils learn; the computer should be for training teachers, training by research is prob-
just a tool to implement new strategies and new ably one of the best. It does not mean that all teach-
solutions to learning problems. It can be an ef- ers should be researchers. But they should be able
ficient tool, for example for individualization of to use the methodology of research, and this can
the learning and for evaluation, but only if indi- be learnt through group activities - reflection, inno-
vidualization or evaluation problems are solved vation, preparation of documents and of situations,
in pedagogical and didactical terms. Technology etc. Teachers will need to learn to work in teams
does not replace pedagogy. So, training in new with colleagues. To be prepared for such activities,
technologies cannot be independent of training they need team activities in their training!
about pedagogy and didactics. Teachers should also be trained to communicate,
. Didactical engineering. Teachers have to elabo- to read, to write (for their pupils; for their col-
rate the situations needed for pupils. Since they leagues; for publication) since this will also be a
have a large number of tools at their disposal component of their job.
and a large number of choices in terms of strat- Teachers should be prepared for a diversity of
egy, a teacher needs to have the characteristics of pedagogy. There exist many different pedagogical
a didactical engineer, i.e. they must have the strategies, many different pedagogical styles. Too
ability to use the results of research or theoret- often, one is convinced that one of these methods
ical statements and transform them into usable is the best. But it is better to be able to deter-
products. mine, in given conditions, at a given moment, with
given pupils what is the appropriate method to en-
Which methodology for training? able them to learn a specific topic. This implies
The methods used in teachers training are at that in the training itself many different methods
least as important as the contents of the training. and strategies will be used - lecturing with one
It is well known that teachers usually teach, not as computer in the room, used mainly by the teacher
they were taught to do, but by reproducing the way (blackboard computer); collective activities in a
they were taught. If you only use lectures to train room with one computer for each student or group
teachers (even if you lecture about active methods of students; individual activities on computers; self-
for teaching), they will then mainly give lectures to evaluation using computers etc.
their pupils. Diverse software must also be used in teacher
Teacher Education and Training 91
training - utilities, basic tools such as word- mathematics concepts with the help of comput-
processors, languages, tutorials, open-ended soft- ers. Every year in the course we choose different
ware, multi-media tools etc. Giving teachers access mathematics topics in the curriculum of univer-
to the maximum of diversity increases their freedom sity studies (not in the curriculum of secondary
in their own professional activities. schools because we want the students to be able
to accomplish by themselves the transfer of these
activities to the field of secondary school math-
Which contents?
ematics).
This is again a question without a definite an- l Pedagogical and didactical viewpoint. In this
swer. Of course, the answer should not be the same part, we use and analyse various existing tools
in pre- and in-service training. But, in fact, the (software, textbooks with computer studies in-
topic of computers is new for most in-service teach- tegrated into them etc.). We try to combine
ers, and they need training which is close to that in the fundamental notions of pedagogy and didac-
pre-service. tics of mathematics together with technology.
We can hope that in the future students will have We also try to make the students solve teach-
acquired the necessary elements about new tech- ing problems using computers. (For example: I
nologies in their previous studies, for example using must prepare a lesson about linear equations for
computers for word-processing. tomorrow. What will the content of the lesson
The first question to be asked is: Do the teachers be? What software will I use? What will be the
I train need computer science? Will they consider activities of the pupils? Here is another example:
computer science as a subject in itself? Should they I must prepare a course about linear equations,
only learn informatics as it impinges on their main but I have six months to prepare it. How will I
subject, and through its use in their subject? do it?).
Many different answers to these questions have During the year the students have to produce a
been attempted. Some countries have tried to train personal project which takes the form of a piece of
teachers by giving them a full year of training in software they design and experiment with.
computer science. This produces specialists, but
the reinvestment for other teachers was not easy. Research, innovation and training
Many countries organize sessions for teachers. Here
again the diversity of what is offered to the teachers The development of the use of new technologies
is certainly a good thing - lectures on specific topics; in mathematics teaching makes it necessary that re-
one week or two weeks sessions; a course over one search be carried on in several domains - research
term or one year etc. about mathematics learning; research about com-
In pre-service training, there should certainly be puters in mathematics teaching; research leading to
specific modules in order to prepare future teachers appropriate software; research about teacher train-
for the use of computers. As a possible example, ing. This research may take several forms - funda-
here is the contents of a course we have given for mental research, applied and experimental research,
many years at Grenoble University, both to future innovation. Too often, there is a gap between the
teachers and to in- service ones. This course lasts fundamental results of educational research, and
for 150 hours (5 hours a week during 30 weeks). products which are usable in teaching and in train-
The title is: Informatics for mathematics teaching. We need to develop applications and implemen-
ing. Each week, 2 hours are devoted to lectures tations of the results of the research and we need
and 3 hours to practical work. The course is di- pedagogical products based on research.
vided into three parts: Innovation and research can contribute to teach-
l Informatics and algorithmics. Students learn the er training. Indeed, the participation of teachers or
basic use of a computer; they also learn elements future teachers in innovative activities is a good way
of algorithmics, including recursion, proof of pro- for training them.
grams, evaluation of algorithms and data pro- The participation of teachers in elaborating and
cessing. They use three different languages for experimenting with pedagogical products is neces-
programming: Pascal, Logo, and Prolog. sary, but not sufficient. Designing good software
l Mathematics from an informatics viewpoint. In needs computer scientists and software specialists as
order to use the computer in mathematics t.each- well as specialists of pedagogyand teachers (prac-
ing, it is necessary to use it for mathemati- titioners). It is a professional matter which needs
cal activities, and therefore to reconsider some professionals.
Two other tracks need to be explored: at the beginning of a career; training never ends,
Teacher training is becoming more complex, and reflection never ends; in-service training should be
we need courses and training activities adapted considered as a natural component of the teachers
for this purpose. Courses for teachers and future job. We must never forget that teachers are profes-
teachers must be developed. We also need to sionals, and need professional training.
reflect about the specific competencies which a Many countries use a cascade model for
teachers trainer must have. In fact, in many teacher training. The education ministry organises
countries the first problem to be solved before a course for a number of selected trainers; after-
we are able to train ALL the teachers is to train wards, each of them trains a number of other train-
teacher trainers. ers, who then train teachers (or trainers who train
In order to diversify the tools usable in teacher . . . ). Such a model can be efficient, but may also
training, it would certainly be interesting to de- not be! The main characteristics of good training
velop computer tools and software for teacher - motivation, activities, understanding - must be
training. present at each stage of the cascade. And this
model can apply only to very specific, precise, and
limited training.
Conclusion
Training plans have been set up and imple-
We have done a lot of experimenting with new mented in many countries for training teachers in
products and new strategies, and the most enthusi- the use of computers. They have only partly suc-
astic teachers have shown both their efficiency and ceeded. One reason for this is that they are gen-
their limits. The problem now is to generalise the erally too restricted as to technology. A training
use of new technologies, so that ALL teachers are plan should be more global, aiming not only at solv-
able to use them as they wish, or to know why they ing new technology problems, but aiming at solving
do not want to use them. teaching and learning problems. New technology
Two conditions seem to be essential in order to problems should not be treated in too isolated a
help all teachers and future teachers use computers: context.
l Make the computer actually available and us- REFERENCES
able; make it a daily life tool; make it really
user-friendly. This means that it is necessary Ball, D. et al (Eds.) [1987]: Will Mathematics
that schools be well equipped. The aim should Count? Computers in Mathematics Educa-
be that each teacher or future teacher has a com- tion, an AUCBE report.
puter (either one the teacher owns (special plans Bosler, U. and Squires, D. [1989]: Training teachers
may need to be set up for purchasing comput- to design educational software in Educational
ers at reasonable prices; loans may need to be Software at the Secondary Level, (Tinsley, J.
obtained for future teachers) or the institution D. and van Weert, T. J., Eds.), Amsterdam:
must own computers and make them available Elsevier
for teachers and future teachers). [1988]: Pre-service teacher education, Proceedings
l Actually use the new technologies in teacher of ICMEG (Hirst, A. and Hirst, K., Eds.), Bu-
training, and not just train about how to use dapest: J. Bolyai Math. Sot.
them.
[1988]: Information Technology in Computer Edu-
In no case can the technology replace the peda-
cation, IBM Denmark.
gogy. A bad teacher using computers will certainly
still be bad! So training and education are neces- [1990]: The Teacher Today: Tasks, Conditions,
sary, but not only from the viewpoint of technology. Policies, OEDC.
We need coherent training, integrating both techno-
logical and pedagogical approaches. Teachers must
be ready to evolve and adapt, and must retain the
ability to ask questions. At each instant they should
ask whether education or technology is the driving
force.
Teacher training is a continuous process. Pre-
service and in-service training are strongly linked,
and both are necessary. No longer can a teacher be
provided with all the abilities and knowledge needed
THE IMPACT OF SYMBOLIC MATHEMATICAL SYSTEMS ON MATHEMATICS EDUCATION
Bernard R. Hodgson
UniversitC Laval, Quebec, Canada
Eric R. Muller
Brock University, Ontario, Canada
Symbolic manipulators, that is, computer pro- mathematics education. The Appendices provide
grammes with the capability of carrying out sym- the following additional information: (1) references
bolic computations, for example, in calculus or lin- dealing with the technical aspects of some of the
ear algebra, are now widely available. While these better known Symbolic Mathematical Systems, (2)
are well-established tools in many areas of math- further illustrations of the capabilities of these sys-
ematics, science and engineering, it must be recog- tems, and (3) references to current projects aimed
nized that they are still in their infancy with respect at the integration of such systems into mathematics
to their use in mathematics education. They repre- education.
sent an ineluctable challenge to current approaches
1. Symbolic Mathematical Systems
to the teaching of mathematics and there is a belief
among some members of the mathematical commu- The term Symbolic Mathematical Systems is
nity that electronic information technology, through used to define calculator and microcomputer sys-
these symbolic capabilities, will exert a deep influ- tems which provide integrated (1) numeric, (2)
ence on how and what mathematics is taught and graphic, and (3) symbolic manipulation capabili-
learned (for example see Page (1990)). However no ties. Numerical computations have always been
clear pattern has yet emerged on how such an influ- included in the domain of both the calculator and
ence is to be articulated. the computer. This capability is usually thought
This paper will discuss certain aspects of the im- of as the ability of doing decimal arithmetic. For
pact of symbolic manipulators on mathematica,l ed- example, if l/3 + l/9 is input, then the approxi-
ucation in the upper secondary years and the first mate solution 0.444444 (to some prespecified num-
few years of university. It is by no means intended to ber of digits) is provided. Symbolic Mathematical
give the final word on such a vast field as much work Systems have the ability to perform rational arith-
is in progress and the technical environment (com- metic, that is, to give the exact answer 4/9 if the
puter hardware/software and calculators) is con- input is l/3 + l/9. The user must request the dec-
stantly improving. The aim of this paper is rather imal approximation if it is desired. Graphing is a
to examine some of the major issues and to indi- more complicated numerical activity. Calculators
cate general trends which have developed since the with graphic capabilities (for example the Casio
1985 ICMI Study on The Influence of Computers fx-7000G, Hewlett-Packard HP-48SX or Texas In-
and Informatics on Mathematics and its Teaching.
The influence of symbolic manipulators on more ad- It should be noted that in a much more gen-
vanced (senior) mathematics courses will not be ex- eral context, the expression symbolic computa-
plored. This is not intended to belittle their impact tion could be construed as referring to various
at this level but rather to concentrate on those years types of symbolic objects, for example as described
where these systems must be implemented in order by Aspetsberger and Kutzler (1988): geometric ob-
to benefit the largest possible number of students jects (computational geometry), logic objects (auto-
in mathematics courses. The influence of these sys- matic reasoning), programmes (automatic program-
tems and their mathematical foundations (see for ming). The concerns of this paper are limited
example Davenport, Siret and Tournier (1988)) will to computations involving algebraic expressions, so
be thrust into the upper level courses by more capa- that typical topics of the field are symbolic differ-
ble and interested students as they progress through entiation and integration, calculation of sums and
the system. limits in closed form, symbolic solution of syst,ems
Section 1 defines Symbolic Mathematical Sys- of equations and of differential equations, polyno-
tems in broad terms and present.s an example of mial factorization, manipulation of matrices with or
their potential use in mathematics education. Sec- without numeric entries, arbitrary precision rational
tion 2 raises some general concerns related to the arithmetic computations, etc. These are sometimes
impact of these systems on mathematics education misleadingly called Computer Algebra Systems -
while Section 3 discusses implementation of some but they can do much more than algebra as will be
of the required changes in secondary and university illustrated by the examples in this article.
93
struments TI-81) as well as microcomputer graph- ::: evalf(subs(t=4.0OI,y));

ing programmes are available. To many mathe-
maticians and mathematics educators symbol ma- .249969
nipulation by calculators (for example the Hewlett- #Looks as though the function is approaching 0.25
Packard HP-28s or HP-48SX) and microcomputer
#as t approaches 4. Does the graph support this?
programmes (for example Maple, Mathematics, De-
rive to name a few) was a most unexpected de- #A plot of y for 3.5 < t < 4.5 is obtained.
velopment. It is the one capability which has the ::: plot(y,3.5..4.5);
potential of producing the most radical changes in
the teaching of mathematics at the secondary school
and university levels.
To convey a feeling for some of the capabilities
of Symbolic Mathematical Systems and how they
could be used in a calculus class, consider the fol-
lowing example of a session with a specific system
(namely Maple but this particular choice is not cru-
cial). Such an example could be done in class, or
could be structured as part of a laboratory exer-
cise. The example illustrates the numeric, graphic
and symbolic manipulation capabilities of the sys-
tem and shows the system can be used in a mode
which requires no programming by the user, but
only the knowledge of a few command words. For
ease of understanding lines starting with a # (and
in italics) are external comments, lines starting with
a ::: are the users input and the lines in bold are
the (Maple) systems response. #Yes it does and the graph indicates by a hole that
#the function is not defined at t = 4, where y is
#The task is to explore the derivative of in(x) using
#the definition of the derivative. First the limit of #approximately equal to 0.25. One repeats this
# (In(t)-ln(4))/(t-4), called y, as t approaches the #experimentation with a few more integer and
# integer value 4 is explored from a numerical #rational cases, for example 5, 3/Z, Y/3. Then the
#point of view, by computing the value of y around #symbol manipulation capabilities can be used to
#t = 4. Clearly the value at t = 4 does not exist. #evaluate the limit directly,
::: y:= (In(t)-ln(4))/(t-4); ::: limit((ln(t)-ln(3/2))/(t-3/2),t=3/2);
y := In(t) - 144)
t-4
#At t = 3.99
#suggesting that the limit of
::: subs(t=3.99,y);
#(In(t)-ln(a))/(t-a) as t approaches a is l/a
-100 ln(3.99)+100 ln(4) #for all a>O. This is confirmed by the system.
#Evaluation using floating-point arithmetic of this ::: limit((ln(t)-ln(a))/(t-a),t=a);
#last displayed expression then gives
1
::: evalf(); -
a
.250313
#Which zs a1so confirmed by the diflerentiation
#At t = 3.999
::: evalf(subs(t=3,999,y)); #capability of the system.
::: diff(ln(a),a);
.250031
-1
#At t = 4.01 a
::: evalf(subs(t=4.0l,y)); #Does the derivative have the properties expected?
.249688 #Plot the function and its derivative on the same
#At t = 4.001 #graph (in some judiciously chosen interval !).
Symbolic Mathematical Systems and Mathematics Education 95
::: plot({ln(t), l/t}, 0.5..10.5); #range. This is visualited with the following plot,
#where we notice that the graph of y is
#completely contained in the specified window.
::: plot({y,0.24,0.26},3.6963..4.3379);
#Yes in(t) is a monotonically increasing

#function and the derivative is shown to be
#positive in the chosen range. Measurements along
#the axes appear to confirm the previously #To demonstrate how incredibly sensitive and
#computed points (4,1/4) etc. The formal #accurate the limit procedure is, one can consider
#definition of the limit can also be explored, namely, #the following.
#L is the limit of y as t tends to a if for every ::: limit((ln(t)-ln(3.2))/(t-3.2),t=3.2);
#eps > 0 there is a de1 > 0 such that if
undefined
#0 < It-al < de1 then Iy- LI < eps.
#Consider the case explored earlier where it was #What happened? To resolve this apparent
#conjectured that the limit of y as t tends to 4 is #anomaly the user must realize that elementary
#l/4. Select eps = 0.01 > 0; is there a de1 such #functions involving numbers other than integers
#that for 0 < It - 4 1 < del, then Iy - l/41 < #or rationals are approximated (the calculator
#O.Ol? The condition ly- l/41 < 0.01 can be #mode), that is ln(s.2) is evaluated as shown by
#rewritten as l/4-0.01 < y < l/4+0.01. #the following output
#This is solved using the system. (Maple solves ::: y:=(ln(t)-ln(3.2))/(t-3.2);
#dinerent types of equations: algebraic, numeric, In(t) - 1.163150810
#diflerential, etc.) In this case we are y :=
t - 3.2
#interested in the numerical solution of an
#and, because of the numerical approximation, the
#equation in one variable. Numerical procedures
#limit of y as t approaches 3.2 does not exist.
#for the solution of such equations often require
#the user to specify an interval within which one For those who are not familiar with Symbolic
#expects to locate the root. In this particular case Mathematical Systems Appendix 2 provides further
#Maple does not require such a prompt and examples of their capabilities. While special pur-
#provides the following: pose packages have been created to cover specific
::: fsolve(y=0.24,t); aspects or topics within the mathematics curricu-
lum, this paper is concerned with full service Sym-
4.33789986
bolic Mathematical Systems which can become part
::: fsolve(y=0.26,t); of mathematics education across different courses
and at different levels. The more powerful systems
3.696303966
were originally created to help individuals perform
#From these two values it zs concluded (based on complicated yet algebraically routine mathematics.
#the continuity of the log function) that there is a There is no evidence that the introduction of inex-
#del, for example 0.2, such that when perienced students to more dedicated (smaller spe-
#0 < It - 41 < 0.2, then. y is zn the specified cially developed systems addressing one part of the
syllabus) has been more successful than introducing different roles played by the left and right hemi-
them to the larger more sophisticated systems. spheres of the brain, it is most likely that the repre-
sentation of mathematical concepts in complemen-
2. Mathematics Education Concerns tary modes such as numeric, graphic, and symbolic
will enhance the learning process. For the first time
Mathematics educators must continually make
in the history of mathematics education Symbolic
decisions about what mathematics is to be taught,
Mathematical Systems offer the ability to move ea.+
how it is to be presented and what student activi-
ily and rapidly between these different representa-
ties are to be required or encouraged. To this de-
tions. It is expected that the use of paper and pen-
cision making must now be added the role of Sym-
cil will be retained by most students; however, one
bolic Mathematical Systems. These systems are a
should not be surprised to find students who can op-
fact of life and can no longer be ignored. Mathe-
erate completely within the computer environment
matics educators have the responsibility to decide
since most systems now provide for easy interplay
consciously whether this environment is to be in-
between word processing and Symbolic Mathemat-
cluded within the students educational experience
ical Systems.
and what should be the exact role of the Sym-
bolic Mathematical System. This decision cannot b) Implications to the teacher
be taken lightly for these systems can perform all the For the teacher Symbolic Mathematical Systems
mathematical techniques presently included in sec- are remarkable not only because they can be used to
ondary school mathematics programmes and most directly perform rational, symbolic or graphic com-
of those included in the first two years of university putations but, more importantly, because of what
mathematics. The decision to include or exclude the they suggest about mathematics itself and about
experience of a Symbolic Mathematical System has mathematics teaching. As Young (1986) puts it,
far reaching implications to the student, the teacher (...) we are participating in a revolution in math-
and to the curriculum. These are now considered in ematics as profound as the introduction of Arabic
turn. numerals into Europe, or the invention of the cal-
a) Implications to the student culus. Those earlier revolutions had common fea-
tures: hard problems became easy, and solvable not
The magnitude of the experiences promised to only by an intellectual elite but by a multitude of
the student by Symbolic Mathematical Systems is people without special mathematical talents; prob-
illustrated by the following allegory: lems arose that had not been previously visualized,
A person explores her surroundings by walk- and their solutions changed the entire level of the
ing (pencil and paper) - many interesting things field. Symbolic Mathematical Systems are part
are discovered, but situations in the neighbouring of this revolution. They can serve to help concept
province are too far away to be experienced, so the development and, by permitting easy and efficient
use of a car (standard scientific calculator) is al- processing of non-trivial examples, they can stim-
lowed. As she drives along, local attractions are ulate exploration and search for patterns2, general-
overlooked in order to get to her destination. How- izations or counter-examples. The teacher must now
ever, even with this mode of transport, she can- question the whole of mathematics education. For
not explore distant lands, so an airplane (Symbolic example, it is increasingly difficult to justify want-
Mathematical System) is provided. She lands in a ing students to become good symbol manipulators
country where the language is not her own, customs unless it can be shown that, such procedural skills
are different - as educators we would try to pre- are essential to an understanding of the underly-
pare her for this shock - but there is nothing that ing mathematical concepts - but no one has yet so
is quite like being there. What potential benefit shown. However this does not imply that students
awaits her! - she can now explore concepts which
were unknown before and she can contrast, corn-- The rapid growth of computing and applica-
pare and have a different view and appreciation of tions has helped cross-fertilize the mathematical sci-
her own culture and home environment. In this new ences, yielding an unprecedented abundance of new
land she continues to use the other modes of trans- methods, theories and models. (. ) No longer just
portation, namely, walking and driving to enhance the study of number and space, mathematical sci-
her experience. ence has become the science of patterns, with the-
Mathematics education has many of the proper- ory built on relations among patterns and on ap-
ties of this allegory. Individuals develop their math- plications derived from the fit between pattern and
ematical understanding in various ways. Due to the observation. Steen (1988).
no longer need develop symbol sense -just as the and not by viewing it. Unlike film, Symbolic Math-
arithmetic calculator has not reduced the need for ematical Systems provide an active environment re-
number sense. Suddenly the teacher is brought quiring constant intervention and change of direc-
to question both the content of the mathematics tion. Nevertheless it would be naive not to realize
courses and their presentation. As the former re- that many teachers will find the sacrifice of tradi-
lates to curriculum more directly, the latter concern tional security quite threatening. This will be es-
is addressed first. pecially true of mathematics teachers who see their
The teacher must consider many factors which role as one of professing well-polished mathemati-
affect the learning of mathematics. An important cal knowledge. White (1989) has suggested that the
factor is the social environment. Some students find use of Symbolic Mathematical Systems can be as-
it easier and more enjoyable to work on their own similated most easily in traditional teaching meth-
while others prefer to work in groups. Some de- ods and curricula. However, in practice, finding
pend on the verbal or written or visual presenta- an appropriate role for the teacher may prove to
tion of mathematical concepts by those who under- be a major barrier for the universal introduction of
stand them. Others find this distracting and prefer Symbolic Mathematical Systems into the traditional
to work directly from books. Computers provide lecture presentation and teachers should seriously
opportunities to enhance these social environments. look at alternative and/or complementary modes of
They also introduce a new factor -- the computer implementation. Even though introducing an occa-
- which may, for some individuals, erect new bar- sional Symbolic Mathematical System demonstra-
riers and difficulties. It is therefore important for tion into a traditional set of lectures is a start, what
mathematics educators to provide alternative en- is needed is a complete rethinking of the objectives
vironments for students to experience. Individuals of those lectures.
will then be in a position to evaluate them and de-
(2) The technology can also be used in scheduled
cide which provide the most opportunities for the
laboratory sessions. This is probably the least
development of their mathematical knowledge.
threatening mode of introduction for the teacher.
Symbolic Mathematical Systems can be inte-
Laboratory activities can be developed and tested
grated into mathematics education in a number of
before the students try them. Students can be given
different ways. The three most obvious ones a.re:
materials to prepare for the laboratory sessions and
(1) The teacher can use it as part of a lecture or support can be provided for the students during
class presentation. This requires some projection their scheduled laboratories. The physical labora-
facilities to allow the students to see what appears tory setup can vary. There are advantages to having
on the computer screen. For the mathematics in- students working with their own system and advan-
structor the use of such a system in the classroom tages to having four to six students working together
provides very different class dynamics. Attention with a single system. Activities appropriate for lab-
has to be paid to typing, errors, unexpected forms of oratory work with a Symbolic Mathematical Sys-
expressions, graphs which appear different from the tem should not be a simple duplication of activities
traditional book presentation (cf. Muller (1992)), which can be achieved just as easily with pencil and
multiple answers, etc. Many mathematics instruc- paper. What are appropriate activities? Clearly
tors find this situation difficult to handle. Perhaps the lack of sustained experience limits ones vision.
the central aspect in the successful integration of a Nevertheless it is suggested (cf. Muller (1991)) that
Symbolic Mathematical System in the classroom is laboratory activities should meet one or more of the
a necessary evolution of the role of the teacher where following general attributes:
intervention is no longer restricted t,o exposition. In- (a) they encourage exploration of mathematical con-
stead the teacher must become a facilitator cre- cepts;
ating a context appropriate for a fruitful interaction
(h) they probe inductive reasoning and/or pattern
between the student, the machine and the mathe-
recognition;
matical concept. The lecture-examples format must
(c) they investigate interrelationships between dif-
be replaced by a more open-ended approach. Al-
ferent representations - algebraic, graphical, nu-
though such a point of view is desirable even in a
merical, etc.;
computer-free classroom, it becomes essential when
computers come into play. One of the reasons why (d) they involve problems which would be very diffi-
films and videos have played such a small role in the cult and/or too time consuming to solve without
mathematics classroom may be the mathematicians the technology.
belief that you understand mathematics by doing it One can visualize a situation where the lecture and
laboratory activities are merged and the lecture pre- of students being considered and their level. For in-
sentation takes place in an area where students have stance, one could have requirements for a student in
access to systems. Because students work at differ- a university mathematics programme different from
ent rates with systems it is quite a challenge to lec- those for a student registered in a mathematics ser-
ture in the traditional way and have students work- vice course. In this respect there is much evidence
ing independently or in groups. The lecture dynam- that shows that scientists from other disciplines (see
ics parallels the situation where one allows time for for example Lance et al. (1986)) serviced by math-
students to work independently on problems. De- ematics departments are interested that their stu-
vitt (1990) and others have used this method. dents not be denied the use of Symbolic Mathe-
matical Systems. Such scientists, often more open-
(3) It is important to prepare for the time when stu-
minded than pure mathematicians with respect to
dents will have easy individual access to Symbolic
technological developments, simply perceive Sym-
Mathematical Systems. A consequence of techne
bolic Mathematical Systems as tools that can help
logical improvements is that a calculator with inte-
them in their work and so are eager to use them. It
grated numeric, symbolic and graphic capabilities is
is therefore necessary to reassess the proper balance
no longer a dream and that such devices can only
in the requirement of basic symbolic manipulation
become progressively more powerful and cheaper.
skills and in the choice of topics covered in the var-
Furthermore, one can expect that the difference be-
ious mathematics curricula.
tween portable computers and calculators will be-
come less apparent. Denying the use of such calcula- Mathematics educators must make sure that in
tors/computers in structured mathematics instruc- connection with domains where Symbolic Mathe-
tion does not solve the problem of their existence, matical Systems can play a role, their courses help
and their access by a few more fortunate students. students acquire the appropriate intellectual skills.
Every society believes that its students should be The required skills, while not really new, are very
exposed to all environments which promise a richer often given little place in most traditional teaching:
educational experience. Of course many situations these are interpretive skills, needed to make math-
arise where that society cannot afford to provide a ematical judgements, to appreciate the validity and
particular environment. Nevertheless this does not limitations of the tool being used, to assess the rea-
relieve teachers from their responsibility to make ev- sonableness of the computed answer (cf. Hodgson
ery possible effort to provide them. (1990)). Such skills, being much more demanding
than traditional algorithmic ones, will require the
c) Implications to the curriculum student to be confronted with a substantial number
There is no doubt that Symbolic Mathematical of theoretical notions. Thorough understanding of
Systems will have impact on the curriculum. What mathematical concepts is thus now surely as - or
is in question is the magnitude of this impact. There even more - necessary in mathematics education
is already evidence that traditional courses will have as it has ever been (cf. Hodgson (1987)).
to change if these systems are to be integrated in any Another issue which is important in a pedagogi-
meaningful way. Even with relation to elementary cal context is the extent to which the symbolic pack-
concepts such as graphing, Dick and Musser (1990) age will act in a black box mode or on the contrary
observe: This change in approach made possible give indications about how the answer to a partic-
by these calculators marks a significant shift in how ular problem can be obtained. A White-Box/Black-
graphing could be perceived by students. Instead of Box Principle has been advocated by Buchberger
as a final task to be completed, graphing can assume (1990) in relation to the question: Should students
the role as a problem-solving heuristic and a tool learn integration rules? Buchbergers point of view
for exploration. Thus the traditional calculus ap- is essentially that in a stage where a certain math-
proach of finding what the graph looks like is turned ematical topic is being learned by the student, the
around to using calculus and numerical methods for use of a Symbolic Mathematical System realizing
locating more accurately the properties which are the pertinent algorithms as black boxes would be
known to exist. Students rapidly come to appreci- a disaster. So he calls for systems that would fea-
ate both the exactness of non-numerical algebra and ture the possibility to use an algorithm both as a
the approximation techniques underlying numerical black box (as is most often the case with exist-
analysis. ing systems) and as a white box, i.e. in a step-
The decision as to what extent Symbolic Math- by-step mode in which the reduction of the prob-
ematical Systems are to be included in the mathe- lem to subproblems is exhibited and in which the
matics curriculum will vary according to the groups user could eventually interfere. A similar view is
taken by Mascarello and Winkelmann (1992) in this cal Systems is expanding rapidly. Some of them
volume. They claim that even if all the details of have even been developed specifically for education
the internal functioning of a Symbolic Mathematical at secondary school or at the beginning of univer-
System are not usually essential to users, they must sity education. In most systems, especially the more
not remain totally hidden: understanding of main recent ones, attention is being paid to make them
ideas and fundamental restrictions are necessary for more user friendly, that is, easier to use. A list of ref-
proper use of (what they now call) the grey boxes. erences which review some of the better known sys-
Once these systems have been introduced into tems is provided in Appendix 1. While the choice of
the mathematics courses then the student evalua- a specific Symbolic Mathematical System appropri-
tion must change to reflect this new environment. ate for use in a given classroom context might rest
As less emphasis is placed on certain techniques and on various criteria (e.g. hardware facilities, level of
more time is spent on concepts, the testing proce- instruction, topics to be covered, etc.), it is clear
dures must also change. Osborne (1990), Beckmann that some basic requirements must be met by those
(1991) and others have started to address this issue. systems. For instance the use of the software should
It is clear that much experimentation and re- be transparent, that is students should spend their
search are needed to establish how best to use Sym- time thinking about the mathematics, and not how
bolic Mathematical Systems in the different courses, to operate the computer. Documentation should
with the wide-ranging mathematical capabilities of be essentially unnecessary for users, so that what
students, and with the various attitudes of teachers. needs to be done at any point should be apparent
Appendix 3 provides a list of ongoing projects which (some on-line help facility might however be use-
are addressing some of these concerns. ful in this respect). The software should be robust so
that students (sometimes unpredictable) behaviour
3. Effecting curriculum changes should not cause it to crash or hang up too easily. It
should interact easily with some word-processor, ei-
Generally curriculum changes in the secondary
ther internally to allow preparation of notebooks
school system require much time to be implemented
integrating word-processed text inserted in the mid-
but when they happen, they are universally imple-
dle of active symbolic software code, or externally
mented: this is a direct consequence of the highly
to facilitate preparation of reports by students. But
centralized administration of secondary school pro-
most important of all the program, whether used in
grammes in almost all educational systems. On
a tutorial or interactive mode, should be devised so
the other hand curriculum changes in university
as not to foster the myth of computer omniscience
courses can be far more spontaneous, but they tend
and infallibility too often rooted in students minds:
to be localized to a particular course or section of
while the computer brings in speed and reliability,
a course, usually under the commitment of one or
it is the human being who has the intelligence and
a few highly motivated individuals. Therefore the
the ability to reason and make decisions.
introduction of Symbolic Mathematical Systems in
secondary school and university mathematics edu- As the cost of basic microcomputer technology
cation poses problems of a different nature. In the continues to drop, one would hope for an analo-
former, to affect curriculum change one must con- gous reduction in the price of hardware necessary for
vince a small group of influential curriculum mak- supporting Symbolic Mathematical Systems. While
ers. For the latter, to ensure that the use of Sym- this has happened in some cases, this is not the
bolic Mathematics Systems becomes integrated in general rule. Indeed, one should be aware that
courses, it is necessary to expose the majority of fac- the general software development trend has been
ulty members of the department to these syst,ems. to demand more and more memory and disk space,
Kozmas (1985) study on instructional innovation thereby requiring more powerful and more expensive
in higher education supports this view. He con- microcomputer units. Software developers tend to
trasted projects which were collaboratively devel- think in terms of the latest available (or forthcom-
oped with those developed by individuals and found ing) hardware facilities, and experienced users call
that the former were much more likely to be institu- for more integration, namely word-processing, sym-
tionalized. This section discusses some of the time bol and graphic manipulation, spreadsheets, etc.,
and effort consuming activities which are required all of which push up the requirements of the com-
when introducing Symbolic Mathematical Systems puter system. Thus the implementation of curricu-
in both upper secondary and university mathemat- lum change involving Symbolic, Mathematical Sys-
ics education. tems requires financial planning for the purchase of
The number of different Symbolic Mathemati- equipment and software. Budgets must also be al-
100 InAuence of Computers and Informatics on Mathematics and Its Teaching
located to the maintenance of both hardware and not as positive as what could have been expected
software. Mathematics departments generally have (see for instance Muller (1991) for an attitudinal
little experience in requesting monies. This has survey of some teaching experience with a Symbolic
tended to be the prerogative of Science, Physical Mathematical System).
Education, Fine Arts and other departments, In The principal word of warning is certainly that
secondary schools funds are sometimes allocated for implementing the necessary curriculum changes
implementation of curriculum changes, but these are takes a lot of human resources in the form of time
unlikely to be sufficient. In a university setting it and dedication. It takes time to conceive the new
may be worthwhile to run experimental sections to course, to develop meaningful students activities,
.accumulate evidence of improvements in traditional to prepare new materials, to devise tools for assess-
indicators and to obtain faculty and student atti- ment. And this must be done in contexts where of-
tudes and responses to these systems. ten no (or little) credit is given to those who embark
In a context where a lot of importance is given on such a task! Furthermore released time, super-
in the literature to various symbolic software run- vision, hardware and software all require financial
ning on microcomputers, it might be tempting to resources in an area where administrators have not
overlook calculator technology. But the calculator been used to allocating funds. Mathematics educa-
is not restricted to school applications or to compu- tors must convince school or university administra-
tations on numbers in so-called scientific notation! tions and funding bodies that such an investment is
There are a number of calculator projects reported essential and is worth its value! And what is needed
in the literature (see for example Nievergelt (1987) to support the argument is a critical analysis of con-
and Demana and Waits (1990)). It is true though trolled experiments, rather than anecdotal reporting
that present calculators only have limited graphic of experiences.
and symbolic manipulation capabilities. But devel-
opments in electronic technology strongly suggest 4. Conclusion
that such more powerful and user-friendly calcula- The introduction of Symbolic Mathematical Sys-
tors will most certainly be a reality in a not too dis- tems into mathematics programmes should be con-
tant future. To equip a class or for individual use, sidered within the broader context of the impact of
calculator technology should thus be seriously con- technology on mathematics education. Mathemat-
sidered. This is especially true in situations where, ics teachers who have successfully integrated other
for instance, electricity supplies tend to be unreli- software into their teaching of geometry, statistics
able. etc. as well as computer scientists can offer useful
Once the equipment (hardware/software) has insights and pedagogical points of view. Most of the
been purchased, meaningful mat,hematics activities projects aimed at the integration of Symbolic Math-
for the students must be developed. Few such ac- ematical Systems into mathematics teaching are ei-
tivities are available, although some recent publica- ther still under way or, if concluded, have results
tions provide examples in calculus: see for example which are difficult to interpret. For example, how
the Mathematical Association of America Notes Se- does one separate the effects of a Symbolic Math-
ries (P8) and the Maple Workbook (Geddes et al. ematical System from other effects, such as those
(1988)) referenced in the Bibliography. But redefin- generated by the enthusiam of those involved with
ing objectives for a course or building pertinent ac- the experiment or the effects produced by the avail-
tivities is a daunting task. And for such a quest to ability of additional resources? It is most proba-
have a lasting effect, it should be undertaken not bly too early to look for a significant impact on the
by one individual (with eventual loss of the effect, curriculum (measured by the proportion of students
should that individual be away for a while), but in mathematics courses affected by the existence of
rather by a group, for instance by a majority of Symbolic Mathematical Systems). It appears to be
the faculty members within a mathematics depart- the consensus of those who are using these systems
ment. This raises the difficult question of how to in their teaching that the course is taught differently
react to a possible lack of interest by some of those but that it retains a fairly traditional content.
faculty members. After all, most are busy people Thus there are few proposals of changes in
and are not willing to invest large amounts of their the curriculum narrowly defined by course content,
limited time unless there is some evidence that the Some examples of proposals for change are: Tall
result will be worthwhile. This is even more true (1985,1991) proposes a much greater visual com-
when students attitudes towards the use of Sym- ponent to calculus teaching; Miiller (1990) sug-
bolic Mathematical Systems in the classroom are gests that the conceptual approach to caiculus using
Lipschitz-restricted concepts of limit, continuity, and university mathematics education in the near
differentiation and integration is a much more nat- future. They offer unprecedented opportunities to
ural one for students and one in which Symbolic deepen and revitalize mathematics courses, focus-
Mathematical Systems are easily integrated; Heid ing more on concepts and ideas than on mechan-
(1988) reports experiments in the resequencing of ical calculations. While it is true that Symbolic
skills in introductory algebra and calculus where at- Mathematical Systems, whether on microcomput-
tention to hand manipulation skills was drastically ers or on hand-held calculators, can only become
reduced; Artigue et al. (1988) traces the influence more powerful, more user-friendly and more widely
of computers on the evolution of the teaching of dif- available, they offer right now an exceptional po-
ferential equations; the texts of Hubbard and West tential for progress in the teaching of mathematics
(1991) and of Kocak (1989) support this evolution and there is no reason for mathematics educators to
and emphasize the importance of visualization in delay becoming seriously involved with them. For
the study of differential equations. such an evolution to happen, experiments must be
It is anticipated that many more such experi- performed on a very large scale and results must be
ments will be reported in the near future as there evaluated and widely disseminated.
are many brojects on the way. Appendix 3 lists Appendix 3 contains a (partial) list of projects
some of these projects for which information could presently underway, in which Symbolic Mathemat-
be found. Ralston has constantly advocated cur- ical Systems are being used in the classroom both
riculum reform at all levels of Mathematics Educa- at university and secondary school level. Hopefully
tion in order to reflect the reality of todays tech- these projects can stimulate more mathematics ed-
nology and prepare individuals for future technol- ucators to involve Symbolic Mathematical Systems
ogy; in Ralston (1990), he proposes a framework for in their daily teaching.
the school mathematics curriculum in 2000 which
is highly dependent on the use of technology. Yet Bibliography
teachers receive their mathematics education from
Since the present volume updates the work and
university mathematics courses in which they make
publications of the 1985 ICMI Study on The Influ-
very little use (if any!) of technology. How then
ence of Computers and Informatics on Mathematics
can they be expected to realize the importance of
and its Teaching held in Strasbourg, this dibliogra-
technology in Mathematics Education? The reform
phy is restricted to references which have appeared
must be spearheaded by the universities where t,here
since that meeting.
exists a greater latitude for experimentation.
There is as yet little evidence that Symbolic a) Proceedings
Mathematical Systems have had a significant im-
There are a number of conference proceedings,
pact on the mathematics curriculum of secondary
books of invited papers and series which provide
schools and universities. It appears that the domi-
an overview of classroom and/or laboratory projects
nant reason for this lack of impetus on the curricu-
and raise philosophical and cognitive issues of using
lum is the education of teachers and faculty, that
a Symbolic Mathematical System in mathematics
is, the lack of experience in these systems by a large
education. References Pl and P2 are the outcomes
proportion of mathematicians. In the university set-
of the 1985 ICMI Study.
ting there is no evidence to suggest that changes im-
plemented by an individual in one section of a course Pl) The Influence of Computers and Informatics
will have any impact on the course as a whole un- on Mathematics and its Teaching. Support-
less special effort is directed toward involving the ing Papers of the ICMI Symposium, IREM,
majority of the faculty in a department. There are Universitg Louis-Pasteur, Strasbourg, 1985.
too many interests riding on the required introduc-
tory mathematics courses to expect that innova,tive P2) Howson, A.G. and Kahane, J.-P. (eds.), The
changes made by one individual will be able to per- Influence of Computers and Informatics on
meate the programme without the support from the Mathematics and its Teaching. (Proceedings
majority of individuals in that department. of the ICMI Symposium. Strasbourg, 1985).
Cambridge University Press, 1986.
In spite of the human and financial costs in-
volved, there is no doubt that Symbolic Mathemati- P3) Johnson, D.C. and Lovis, F. (eds.), Informatics
cal Systems must be introduced into t,he mathemat- and the Teaching of Mathematics. (Proceed-
ics curriculum. They probably constitute the single ings of the IFIP TC 3/WG 3.1 Working Con-
most powerful force compelling change in secondary ference. Sofia, 1987). North-Holland, 1987.
P4) Banchoff, T.F. et al. (eds.) Educational Aspetsberger, K. and Kutzler, B. [1988]: Symbolic
Computing in Mathematics. (Proceedings of computation - A new chance for education
ECM/87, Rome, 1987). North-Holland, 1988. in F. Lovis and E.D. Tagg (eds.) Computers
in Education, 331-336, Amsterdam: North-
P5)Demana, F., Waits, B.K. and Harvey, J. (eds.)
Holland.
Proceedings of the Annual Conference on
Technology in Collegiate Mathematics. (1st: Aspetsberger, K. and Kutzler, B. [1989]: Using a
1988; 2nd: 1989; 3rd: 1990). Addison-Wesley, computer algebra system at an Austrian high
1990, 1991. school in J.H. Collins et al. (eds.) Rroceed-
ings of the Sixth International Conference on
P6)Cooney, T.J. and Hirsch, C.R. (eds.) Teaching
Technology and Education, CEP Consultants
and Learning Mathematics in the 1990s. (1990
Ltd., vol. 2, 476-479.
Yearbook). National Council of Teachers of
Mathematics, 1990. Auer, J.W. [1991]: Maple Solutions Manual for
Linear Algebra with Applications, Englewood
P7)Dubinsky, E. and Fraser, R. (eds.) Computers Cliffs, NJ: Prentice-Hall.
and the Teaching of Mathematics: A World
View. (Selected papers from ICME-6, Bu- Ayers, T., Davis, G., Dubinsky, E. and Lewin, P.
dapest, 1988). Shell Centre for Mathematical [1988]: C om p u t er experiences in learning com-
Education, University of Nottingham, 1990. position of functions, J. for Res. in Math. Ed.,
19, 246-259.
P8)The Mathematical Association of America has
issued three volumes in the MAA Notes series Beckmann, C.E. [1991]: Appropriate exam ques-
related to this field and is preparing a fourth tions for a technology-enhanced Calculus I
one: a) Smith, D.A. et al. (eds.), Comput- course in Reference P5 above (1989 Confer-
ers and Mathematics: The Use of Comput- ence), 118-121.
ers in Undergraduate Instruction. MAA Notes
Beilby, M., Bowman, A. and Bishop, P. [1991]:
Number 9, 1988. b) Tucker, T.W. (ed.), Prim-
Maths & Stats Guide to Software for Teach-
ing the Calculus Pump: Innovations and Re-
ing (2nd edition), CT1 Centre for Mathemat-
sources. MAA Notes Number 17, 1990. c)
ics and Statistics, University of Birmingham,
Leinbach, L.C. et al. (eds.), The Laboratory
UK.
Approach to Teaching Calculus. MAA Notes
Number 20, 1991. d) Computer Algebra Sys- Bjork, L.-E. [1987]: Math ematics and the new tools
tems in Undergraduate Mathematics Educa- in Reference P3 above, 109-115.
tion. To appear.
Bloom, L.M., Comber, G.A. and Cross, J.M. [1986]:
P9) The Notices of the American Mathematical So- Use of the microcomputer to teach the trans-
ciety feature a regular column under the title formational approach to graphing functions,
Computers and Mathematics (past editor: J. Int. J. of Math. Ed., 17, 115-123.
Barwise; current editor: K. Devlin).
Brown, D., Porta, H. and Uhl, J.J. [1990]: Calculus
b) Author Bibliography & Mathematics: Courseware for the Nineties,
Mathematics J. 1 (l), 43-50.
In addition to the references mentioned in the
text, the following list contains a selection of some Brown, D., Porta, H. and Uhl, J.J. [199-l: Calcu-
useful papers or books. lus & Mathematics, Reading, MA: Addison-
Wesley.
Adickes, M.D., Rucker, R.H., Anderson, M.R. and
Moor, W.C. [1991]: Structuring tutorials us- Buchberger, B. [1990]: Should students learn inte-
ing Mathematics: Educational theory and gration rules?, SIGSAM Bull., 24, 10-17.
practice, Mathematics J., 1 (3), 86-91.
Capuzzo Dolcetta, I., Emmer, M., Falcone, M.
Akritas, A.G. [1989]: Elements of Computer Alge- and Finzi Vita, S. [1988]: The laboratory of
bra, New York: John Wiley. mathematics: Computers as an instrument for
teaching calculus in Reference P4 above, 175-
Artigue, M., Gautheron, V. and Sentenac, P. (19881:
186.
Qualitative study of differential equations: Re-
sults of some experiments with microcomput- Cromer, T. [1988]: Linear algebra using muMATH,
ers in Reference P4 above, 135-143. Collegiate Microcomputer, 6, 261-268.
Davenport, J.H., Siret, Y. and Tournier, E. [1988]: Hosack, J. [1988]: Computer algebra systems in Ref-
Computer Algebra: Systems and Algorithms erence P8a above, 35-42.
for
Algebraic Computation, Academic Press. Hubbard, J.H. and West, B.H. [1991]: Diflerential
Dechamps, M. [1988]: A European cooperation on Equations: A Dynamical Systems Approach,
the use of computers in mathematics, in Ref- Part I: Ordinary Differential Equations, New
erence P4 above, 197-209. York: Springer-Verlag.
Demana, F. and Waits, B.K. [1990]: Enhancing Hubbard, J.H. and West, B.H. [1991]: MacMath:
mathematics teaching and learning through A Dynamical Systems Software Package, New
technology in Reference P6 above, 212-222. York: Springer-Verlag.
Devitt, J.S. [1990]: Adapting the Maple computer Kocak, H. [1989]: Diflerential and Diflerence Equa-
algebra system to the mathematics curriculum tions through Computer Experiments (2nd edi-
in Reference P5 above (1988 Conference), 12- tion), New York: Springer-Verlag.
27.
Kozma, R.B. [1985]: A grounded theory of instruc-
Dick, T., and Musser, G.L. [1990]: Symbolic/graphi- tional innovation in higher education, J. of
cal calculators and their impact on secondary Higher Education, 300-319.
level mathematics in Reference P7 above, 129-
Lance, R.H., Rand, R.H. and Moon, F.C. [1986]:
132.
Teaching engineering analysis using symbolic
Dubisch, R.J. [1990]: The tool kit: A notebook sub- algebra and calculus, Eng. Educ., 76, 97-101.
class, Mathematics J., 1 (2), 55-64.
Mascarello, M. and Winkelmann, B. [1992]: Calcu-
Ellis, W., Jr. and Lodi, E. [1989]: Maple for lus teaching and the computer. On the inter-
the Calculus Student, Pacific Grove, CA: play of discrete numerical methods and calcu-
Brooks/Cole. lus in th.e education of users of mathematics,
Fey, J.T. [1989]: Technology and mathematics edu- (in this volume).
cation: A survey of recent developments and
Mathews, J.H. [1989]: Computer symbolic algebra
important problems, Educ. Studies in Math.,
applied to the convergence testing of,infinite
20, 237-272.
series, Collegiate Microcomputer, 7, 171-176.
Flanders, H. [1988]: Teaching calculus as a labora-
Mathews, J.H. [1990]: Teaching Riemann sums us-
tory course in Reference P4 above, 43-48.
ing computer symbolic algebra systems, Col-
Foster, K.R. and Bau, H.H. [1989]: Symbolic manip- lege Math. J., 21, 51-55.
ulation programs for the personal computer,
MSller, 11. [1990]: Elementary analysis with micro-
Science, 243, 679-684.
computers in Reference P7 above, 179-184.
Geddes, K.O., Marshman, B.J., McGee, I.J., Ponzo,
Muller, E.R. [1991]: Maple laboratory in a service
P.J. and Char, B.W. [1988]: Maple - Calculus
calculus course in Reference P8c above, lll-
Workbook, University of Waterloo, Canada.
117.
Gray, T.W. and Glynn, J. [1991]: Exploring Math-
ematics with Mathematics, Reading, MA: Muller, E.R. [1992]: Symbolic mathematics and
Addison-Wesley. statistics software use in calculus and statis-
tics education, Zentralblatt Didaktik Math. (to
Heid, M.K. [1988]: R ese q uencing skills and concepts
appear).
in applied calculus using the computer as a
tool, J. for Res. in Math. Ed., 19, 3-25. Neuwirth, E. [1987]: The impact of computer alge-
bra on the teaching of mathematics in Refer-
Heid, M.K., Sheets, C. and Matras, M.A. [1990]:
ence P3 above, 49-53.
Computer-enhanced Algebra: New roles and
challenges for teachers and students in Ref- Nievergelt, Y. [1987]: The chip with the college ed-
erence P6 above, 194-204. ucation: the HP-28C, Amer. Math. Monthly,
94, 895-902.
Hodgson, B.R. [1987]: Symbolic and numerical com-
putation: The computer as a tool in mathe- Orzech, M. [1988]: Using computers in teaching lin-
matics in Reference P3 above, 55-60. ear algebra in Reference P8a above, 63-67.
Hodgson, B.R. [1990]: Symbolic manipulation sys- Osborne, A. [1990]: Testing, teaching and technol-
tems and the teaching of mathematics in Ref- ogy in Reference P5 above (1988 Conference),
erence P7 above, 59-61. 60-67.
Page, W. [1990]: Computer algebra systems: Issues many of the evaluations do not take into account
and inquiries, Computers Math. Applic., 19, possible classroom use and the use by neophytes.
51-69.
a) The Notices of the American Mathematical Soci-
Ralston, A. [1990]: A framework for the school ety (see reference P9 above) have recently included
mathematics curriculum in 2000 in Reference an individual review of most Symbolic Mathemati-
P7 above, 157-163. cal Systems:
Shumway, R. [1990]: Supercalculators and the cur- Vol. 35, 1988
riculum, For the Learning of Math., 10 (2),
The HP-Z8S brings computations and theory back
2-9. together in the classroom, Y. Nievergelt, 799-
Small, D. and Hosack, J. [1991]: Explorations in 804.
Calculus with Computer Algebra Systems, New Supercalculators on the PC., B. Simon and R.M.
York: McGraw-Hill. Wilson, 978-1001.
Small, D., Hosack, J. and Lane, K. [1986]: Com- Mathematics - A review, E.A. Herman, 1334-
puter algebra systems in undergraduate In- 1344. (Also: Other comments on Mathemat-
struction, Coil. Math. J., 17, 423-433. ica, 1344-1349.)
Steen, L.A. [1988]: Th e science of patterns, Science,
Vol. 36, 1989
240, 611-616.
MicroCalc 4.0, G. Gripenberg, 680.
Tall, D. [1985]: Vzsualizing calculus concepts using
a computer in Reference Pl above, 291-295. The menu with the college education (A review of
Derive), E.L. Grinberg, 838-842.
Tall, D. [1991]: R ecent developments in the use of
computers to visualize and symbolize calculus Milo: The math processor for the Macintosh, R.F.
concepts in Reference P8c above, 15-25. Smith, 987-991.
Wagon, S. [1991]: Mathematics in Action, San Fran- Milo, Sha Xin Wei, 991-995.
cisco: Freeman. PowerMath II, Y. Nagel, 1204-1206.
White, J.E. [1988]: Teaching with CAL: A math- More on PowerMath II, P. Miles, 1206-1207.
ematics teaching and learning environment,
Vol. 37, 1990
College Math. J., 19, 424-443.
Review of PC-Macsyma, Y. Nagel, 11-14.
White, J.E. [1989]: Mathematics teaching and
learning environments come of age: Some new Review of True Basic, Inc. Calculus 3.0, J.R.
solutions to some old problems, Collegiate Mi- Moschovakis, Y. Matsubara, G.B. White, 129-
crocomputer, 7, 203-224. 131.
Young, G. [1986]: Epilogue in R.E. Ewing, K.I. Derive as a precalculus assistant, P. Miles, 275-276.
Gross and C.F. Martin (eds.), The Merging The right stuff, K. Devlin, 417-425.
of Disciplines: New Directions in Pure, Ap-
plied and Computational Mathematics, 213- Almost no stu# in, wrong stuff out, J.D. Child, 425-
426.
214, New York: Springer-Verlag.
Zorn, P. [1987]: Computing in undergraduate math- Four computer mathematical environments, B. Si-
ematics, Notices Amer. Math. SOL, 34, 917- mon, 861-868.
923. Vol. 38, 1991
Zorn, P. [1990]: Algebraic, graphical and numerical Crimes and misdemeanors in the computer algebra
computing in elementary calculus: Report of trade, D.R. Stoutemyer, 778-785.
a project at St. Olaf College in Reference P5
Periodic knots and Maple, C. Livingston, 785-788.
above (1988 Conference), 92-95.
b) Other reviews are:
APPENDIX 1 Symbolic manipulation programs for the personal
This appendix provides a list of some Symbolic computer, K.R. Foster and H.H. Bau, Science,
243, 679-684 (1989).
Mathematical Systems software reviews. It is im-
portant to realize that it is extremely difficult to Derive: A mathematical assistant, E.A. Herman,
evaluate and benchmark this softwa.re. Furthermore Amer. Malh. Monthly, 96, 948-958 (1989).
Mathematics: A system for doing mathematics by #calculations. What can Maple do with that
computer, L.S. Kroll, Amer. Math. Monthly, #number?
96, 855-861 (1989). ::: ifactor(267-1);
Math without tears, C. Seiter, Mat World 8 (l), 159- (761838257287)(193707721)
165 (1990).
#Done in just a fraction of a minute!! (But
Theorist, J. Rizzo, MacUser, 6 (6), 57-59 (1990).
#needless to say it is very easy to give as an
Mathematics: A system for doing mathematics by #input a number that would take three years of
computer, A. Hoenig, Math. Intelligencer, 12 #Sundays for the system to do.)
(2), 69-74 (1990).
#Roots of equations can be found directly.
Theorist, F. Wattenberg, Amer. Math. Monthly, ::: y:= x-3-4*x-2-7*x+10;
98, 455-460 (1991).
y:=x3-4x2 -7x+10
Review of Maple in the teaching of calculus, E.R.
Muller, College Math. J., (to appear). ::: solve(y=O,x);
c) Reviews of software and comments on experi- l,-2,5

ments on their use in teaching can also be found #Even with symbolic coefficients.
in specialized newsletters. Some examples are:
::: z:= a*x^2-2*b*x+c;
Computer Algebra Systems in Education Newsletter
z:=ax'-2bx+c
published by the Department of Mathematics,
Colby College, Waterville, ME 04901, USA. ::: solve(z=O,x);
Maths & Stats published by the CT1 Centre 2 b+2 (b2 -a c)li2
for Mathematics and Statistics (Computer in 112 ,
a
Teaching Initiative), Faculty of Education,
2 b - 2 (b2 -a c)li2
University of Birmingham, Birmingham, B15 112
2TT, UK. a
#The example in Section 1 demonstrated that
Computer-Algebra Rundbrief published by Fach-
#these systems are able to compute limits and to
gruppe 2.2.1 Computer-Algebra der GI, c/o
#differentiate. They can also find Taylors Series,
Dr. F. Schwarz, GMD, Institut Fl, Postfach
#sum jnite and infinite series and integrate
1240, 5205 St. Augustin, Germany.
#and solve differential equations.
APPENDIX 2 ::: taylor(ln(x^2),~=1,4);
This appendix provides a limited number of ex- 2(x-l)- (x- 1)2 + 2/3 (x -1)3 + O((x - 1)4)
amples to illustrate some of the capabilities of Sym- ::: sum(j^2, j=l..m);
bolic Mathematical Systems (the system used here
is Maple but this particular choice is not crucial). l/3 (m+1)3-l/2 (m+1)2+1/6 m+1/6
These systems are so powerful that it is impossible ::: int(x*3*sqrt(x^2-9),x);
to provide a complete overview of their capabilities
in a brief text. -3/5.(x2 - 9p2x2-54/5 (x2 - 9p2
+1/5 (x2 - 9p2x4
#The system can be used to do some elementary
#number theory. For instance the command ifactor ::: deq:= diff(x(t),t)*t^3+x=O;
#returns the prime factorisation of an integer.
::: ifactor( 123456780); deq:=(ix(t)) t3 +x = 0
w w2 (5) (47) (14593) ::: dsolve(deq,x(t));

#With such a toolavailable, it might be tempting to
x(t)= exp(l/2 $)C
#venture into some calculations that are not
#trivial to do either by hand or in a standard #The next few examples are taken from linear
#computer environment. For example the prime #algebra, namely, the solution of linear
#factors of the Mersenne number 267 - 1 were #equations and some properties of matrices and
#given in 1903 by F. Cole. It reportedly took him #vectors.
#three years of Sundays to complete the ::: eqns:= a*x+b*y=e,c*x+d*y=f; vars:= x,y;
eqns:={ax+by=e,cx+dy=f} APPENDIX3
vars:= {x,y} There is as yet no single source which can
::: solve(eqns,vars); provide a comprehensive international listing of
projects in the area of Symbolic Mathematical Sys-
bf-ed af-ce
{x = - Y= 1 tems in Mathematics Education. Therefore, the fol-
ad-cb' ad-cb
lowing list cannot be regarded aa comprehensive:
::: A:= array([[l ,a,a^2l,[l,b,b^2l,[l,c,c^211); 1: The Swedish ADM project (Analysis of the role of
A := array(1..3,1..3, the Computer in Mathematics Teaching); see Bjijrk
(1987).
11, a, a"1
2: The Research Institute for Symbolic Computa-
[l, b, b2] tion at the Johannes Kepler University, Linz, Aus-
11,c,cI> tria.
3: The Computers in Teaching Initiative Centre for
::: det(A);
Mathematics and Statistics (Development of class
bc 2-b2c-ac2+a2c+ab2-a2b work sheets to be used with Derive), see the Maths
::: factor(); & Stats newsletter published by the CT1 Centre,
University of Birmingham, UK.
-(- c+ b)(a- c)(a - b)
4: A European Cooperation on the use of Computers
#(The symbol refers to the previously in Mathematics; see Dechamps (1988).
#displayed expression.) 5: The National Science Foundation (U.S.A.) is
::: al:= array([xl,yl,zl]); a2:= array([x2,y2,z2]); funding a number of different university projects
a3:= array([x3,y3,z3]); specifically directed at integrating Symbolic Math-
ematical Systems into the calculus curriculum. The
al := array(l..3,
following is a selection providing a one line state-
[xl, Yl, zll) ment together with the university and the principal
a2 := array(l..3, investigator.
[x2,Y2,4) Developing a user friendly interface to Maple
and incorporating use of system into teaching cal-
a3 := array(l..3,
culus, Rollins College, Winter Park, FL; Douglas
[x3,Y3,z31) Child.
::: vol:= abs(dotprod(al,crossprod(a2,a3))); Developing new calculus curriculum using Maple
on a VAX, Rensselear Polytechnic Institute, Troy,
vol:=abs(xl(y2 23 -22 y3)
NY; William Boyce.
+ yl(z2 x3-x2 23)
Developing a computerired tutor and computa-
+ zl(x2 y3-y2 x3)) tional aid based on Maple, University of Rhode Is-
::: a:= array([[l3,5],[5,2]]); land, Kingston, RI; Edmund Lamagna.
Developing an electronically delivered course us-
a := array(l..2,1..2,
ing the Notebooks feature of Mathematics, Univer-
[13,51 sity of Illinois, Urbana, IL; Jerry Uhl.
PI 21) Developing a new calculus course emphasing ap-
plications and using Mathematics, University of
::: c:= eigenvals(a);
Iowa, Iowa City, IA; Keith Stroyan.
c :=15/2+1/2 2211j2, 15/2-l/2 2211j2 Developing a laboratory based calculus course us-
#The decimal approximation to these two ing Mathematics, Iowa State University, Ames, IA;
#eigenvalues gives Elgin Johnston.
::: evalf(c[l]); evalf(c[2]); Developing a new calculus course for liberal
arts colleges using Mathematics, Nazareth College,
14.93303438 Rochester, NY; Ronald Jorgensen.
.066965625 Developing calculus as a laboratory course us-
ing MathCad and Derive, Duke University, Durham,
NC; David Smith.
Emphasizing computer graphics using Maple and
emphasizing concepts via programming. in ISETL,
Purdue University, West Lafayette, IN; Ed Dubin-

sky.
Porting the laboratory calculus developed at Duke
over to Mathematics, Bowdoin College, Brunswick,
ME; William Barker.
Collecting, teiting, and desktop publishing the
best materials being developed using Mathematics,
University of Michigan at Dearborn, Dearborn, MI;
David James.
More detailed informations about projects in the
U.S.A. integrating Symbolic Mathematical Systems
in the calculus curriculum can be. found in the re-
ports contained in reference P8b above: Tucker,
T.W. (ed.), P riming the Calculus Pump: Innova-
tions and Resources. Mathematical Association of
America (MAA Notes Number 17), 1990.
-___ . ...^_ ~._l_ll__-._- .--- ..-.-.----._____l_l~

CALCULUS TEACHING AND THE COMPUTER. ON THE INTERPLAY OF DISCRETE
NUMERICAL METHODS AND CALCULUS IN THE EDUCATION OF USERS
OF MATHEMATICS
Maria Mascarello
Politecnico di Torino, I-10129 Torino, Italia
Bernard Winkelmann
Institut fiir Didaktik der Mathematik, D-4800 Bielefeld, Deutschland
1. NEW POSSIBILITIES
the integration into regular classroom teaching will
The computer is a mighty mathematical tool,
still be a problem.
not only for mathematical research, but even more
in the process of applying mathematics and in the The classic situation of the user of mathemat-
process of teaching and learning mathematics. In ics could be described - in a somewhat oversimpli-
the following, we shall concentrate mainly on the fied manner - as a huge amount of passive mathe-
new possibilities which the computer presents in the matical knowledge contained in monographs, hand-
realm of calculus for users and future users of math- books, recipes. Traditionally, this knowledge could
ematics. By a user we mean somebody who is in- only be used by being activated through the active
terested in mathematics merely (or mainly) through mathematical knowledge of the user himself or by
the use of mathematical models (in particular calcu- direct cooperation between the user and a mathe-
lus models) to solve (extra-mathematical) problems. matically more knowledgeable person. In contrast
Future users of mathematics are, for example, en- to this, the mathematical knowledge contained in
gineering students, but even those learning calculus mathematical software can have a far more active
in schools as part of a general education may be character, e.g. in giving advice and help interac-
included under this rubric. tively, offering possibilities for exploratory experi-
ments or answering questions, acting like a mathe-
1.1 New possibilities for the user matical expert system. Even more common numer-
We describe first the changes in the mathemati- ical software, which exists in the form of sophisti-
cal knowledge and habits of the user of mathematics cated procedures, is far more active than the recipes
induced by the availability of sophisticated math- of the old-fashioned handbooks, since in many cases
ematical software to all who have to rely heavily these procedures are in fact polyalgorithms: They
on mathematical problem-solving such as engineers, decide with considerable expertise which particu-
natural scientists, etc. During the past decade we lar algorithm should be invoked, depending on the
have seen the proliferation of mathematical software circumstances2. So the demand for mathematical
systems for personal computers which have become knowledge on the part of the user has changed. The
more powerful and/or more user friendly. By rais- emphasis has shifted from detailed knowledge of the
ing the standards in these two domains, such sys- advantages and disadvantages of specific numerical
tems are now in the hands of a rapidly growing num- methods and of the algorithms themselves to some
ber of users, even if until now (1991) they have not meta-knowledge of the possibilities of numerical al-
yet reached the majority of the teachers of mathe- gorithms in general and their interaction with the
matics, at least at the secondary level. But, if the concrete application situation.
trend continues, not only professional users of math- As an example let us look at the process of the
ematics, but also most students and teachers will solution of ordinary differential equations3. This is
soon have regular access to such systems, However, indeed an example of great importance since such
equations appear in many applications and are at
1 A typical example might be the realm of com- the heart of applicable elementary calculus. So if
puter algebra systems: In the progress from MU- it is possible to master them at a more elementary
MATH to Derive there has been a big gain in user level than hitherto was possible, this could even be
friendliness, allowing the use of the system even by regarded as the most appropriate goal for the teach-
users reluctant to program, but - at the same time - ing of elementary calculus at schools and colleges. In
with a certain loss in functionality, e.g. in the solv- the education of engineers at technical universities
ing of differential equations. On the other hand, the
progress from MUMATH to Mathematics is mostly cf. Rice [1983], e.g. p.291f.
in power, much less in user friendliness. See also the 3 cf. Winkelmann [1984] and the chapter by Tall
chapter by Hodgson and Muller in this book. and West in this book.
108
Calculus Teaching and the Computer 109
or similar institutions, where differential equations What does this mean for the mathematical ed-
have always been part of the calculus sequence, even ucation of the future user? Of course, there is no
beginning calculus could concentrate more on appli- direct way from the mathematical activities of the
cations and so give the student a more realistic and, user to the teaching process; the goal must not be
one hopes, a more motivating start. confused with the means. Understanding and abili-
In the pre-computer age an engineer or scientist ties for complex interpretations can only be built up
who had to handle differential equations was sup- by personal involvement of the student; she has to
posed to have detailed knowledge of diverse meth- do full (but simpler) examples in all the main steps
ods for the analytic solution of various elementary herself, be it by hand-calculating, by using interac-
types, to be able to master complicated analytic- tive symbolic systems or calculators or by program-
algebraic formulas and to carry out lengthy error- ming in some suitable programming language. This
free symbolic and numerical calculations. Now he or seems necessary in order to get an awareness of the
she can use software which has this knowledge and mathematical situations, even if such activities are
ability built in, since it can solve more elementary no longer part of the final application process. And
differential equations than a non-specialist mathe- even if todays sophisticated mathematical software
maticiari cati do4. But in building up the model need not and cannot generally be fully understood
the user still has to understand fully the meaning by the normal user, there must not be totally black
and significance of the diverse quantities (variables) boxes; a principal understanding of simple cases, of
and of their derivatives and to be able to relate main ideas or of fundamental restrictions can be
these to each other in order to set up the differential gained and seems necessary for proper use of the
equation. And to give the details to the computer now grey boxes5.
program, a thorough intuitive understanding of t#he On the other hand it is quite clear that extensive
mathematical meaning of the identifiers which ap- drill in formal calculations, in fluent structured pro-
pear in the modeling equations is needed, be it as gramming or even in the handling of some software
variables, parameters, initial values, names for (yet package cannot be justified in view of the changed
unknown) functions (dependent variables) and so qualifications needed by the user.
on. If an analytic solution exists, the program will 1.2 New possibilities in the teaching-
normally present it as a somewhat confusing lengthy learning process I
expression which must be qualitatively interpret,ed In the field of teaching methods the computer,
to be understood, namely through looking for sim- if it has been loaded with the appropriate software,
pler special cases, for settings of specific parame- will function as a simplifying aid, almost as a su-
ters or initial values, for asymptotic patterns of be- per hand-held calculator which permits the pupil to
haviour, etc. This process is guided by the intended overcome computational obstacles in the treatment
interpretation of the solution in the context of t,he of more complex problems and to handle more re-
application model. If no analytic solution exists, t,he alistic applications, e.g. in dealing with larger ma-
user may give his equation to some ready-made nu- trices, in the numerical solution of differential equa-
merical software. In this case he needs some knowl- tions, or in the symbolic treatment of more com-
edge to make reasonable explorative choices of t,he plicated formulas; this will serve to widen the po-
values of parameters and initial values; there should tential scope of mathematics education in terms of
be some experience with numerical phenomena (pit- content. On the other hand, a computer equipped
falls of computations) and the ability to interpret with appropriate languages and environments can
the numerical and graphical output of t,he computer become an instrument for solving problems in the
and to use this interpretation interactively for new
choices of starting points for the next calculation.
5 Buchberger [1990] gives an argument for a much
In total, there can be observed a specific shift more strict procedure: first, the algorithms of the
in the spectrum of abilities, from precise algorith- software have to be completely understood by the
mic abilities to more complex interpretations, so to student; afterwards he may use the software for all
speak from calculation to meaning, which in a cer- calculations. But Buchberger has the algorithms
tain sense is a reversal of the historical evolution. In of Computer Algebra and mathematical majors or
this process the mathematics to be mastered tends computer science majors in mind; his arguments do
to become intellectually more challenging, but tech- not extend to numerical software and typical future
nically simpler. users of mathematics. See also the somewhat more
detailed discussion in the chapter by Fraser, Klingen
4 cf. Watanabe [1984]. and Winkelmann in this book.
hands of the student (interactive calculating or pro- informatics in the main course of lectures and de-
gramming); in this case, the student tends to under- voting special laboratory sections to calculus at the
stand techniques more at the cognitive level, and no computer.
longer mainly at the level of skill. Beyond that, the
computer, with its possibilities for illustration and 2. THE DISCRETE - CONTINUOUS
symbolization, will provide opportunities for more INTERPLAY
comprehensive and rapid mathematical experiences. 2.1 General considerations
This presents problems and tasks as well as op- Although the role of applications of analysis has
portunities for educators mainly on two levels. On a been changed both by the growing number of disci-
more technical level, there is the necessity to provide plines using mathematical models and by new meth-
more suitable software with strong mathematical ods, particularly the extensive use of computers,
functionality, educationally sound help functions, an understanding of fundamental concepts in which
user interfaces for the inexperienced user, and ac- mathematizations take place remains indispensable.
companying explanations, hints and worked-out ex- Examples are:
l variable quantity, change
amples for teachers. On a more fundamental level,
the problem is to achieve a balance between the l functional dependency
quantitative and qualitative relation of new and old l local rate of change
goals and methods as well as to set up the right 0 average value

trends for future developments. 0 accumulation.
The computer creates new opportunities for in- We shall refrain from discussing here how far tra-
struction in analysis, e.g. ditional mathematics education was able to attain
l numerical and graphical illustration@, the goal of teaching these.
l more complex and more realistic applications, Now it is evident that these central concepts
l a language in which to describe traditional cal- of mathematical applications can be implemented
culus, both by discrete and by continuous conceptual-
izations. Corresponding to such continuous con-
l CAL (computer-aided learning) in its various
cepts as function, differential equation, derivative,
forms.
weighted integral, and integral, are the correspond-
Some traditional motivations for treating con-
ing conceptualizations in discrete analysis, namely:
ceptually exacting analysis in school can, however,
sequence and time series, difference equation, dif-
no longer be maintained. For instance:
ference, arithmetical mean value, and sum. These
l calculations such as finding extreme values or discrete concepts are often technically and almost
areas can be easily done without analysis, always intellectually much simpler than their con-
l practical applications in physics or technology tinuous counterparts.
which used to rely on analytic methods now are rou- In the following we will give some justifications,
tinely done numerically on a computer by discrete which are, in our opinion, crucial in answering the
calculations. question now raised inevitably: Why use the con-
This results in a crisis: The legitimacy of tra- cepts of continuous analysis in teaching at all? In
ditional analysis in school is challenged; educators (a) and (b), we state the problem, conceived as an
will have to make clear to the general public, and epistemological question regarding the role of anal-
the teacher will have to explain to his pupils how ysis in applications and model building, and in (c)
and why the treatment of continuous analysis still and (d) we introduce the argument which solves the
makes sense nowadays. dilemma.
In Section 3 we shall report on some experiments (a) Insuficiency of continuous analysis for ob-
concerning the use of informatic tools in teaching taining concrete numerical results. Let us recall
basic mathematical courses at the Politecnico (Poly- some of the facts: Most integrations cannot be ex-
technics) of Torino, Faculty of Engineering Sciences. ecuted analytically, but only numerically; this is
We emphasize that the choice here has been to all the more true for solving differential equations.
keep the teaching of calculus reasonably traditional, Even tasks as simple as determining the extremes
while at the same time giving some basic notions of of a familiar function like 2 sin 2 require numerical
methods. School mathematics has hitherto confined
6 See Tall [1986], the article by D. Tall in John- itself in a rather unnatural way to problems involv-
son/Levis [1987] and the chapter by Tall and West ing classes of functions which were solvable by ana-
in this book. lytic methods. It has paid dearly for this with heavy
losses in orientation to problems of reality, content enter. Calculus concepts such as limit, derivative,
and relevance. This is particularly true for classi- integral are not to be interpreted in the strict math-
cal university courses in, for example, elementary ematical sense, but they express certain invariances:
differential equations7. The corresponding discrete concepts do not depend
(b) Most concrete models using analysis have a on the step size, provided it is sufficiently small (but
discrete basis. This is first evident in the social sci- yet in the intended scaling domain). This consider-
ences or in population biology, where the quantities ation gives sense to the use of calculus models in
to be modelled are numbers of items or individuals, such typically discrete domains as population dy-
or monetary units, which cannot be subdivided at namics or economics. But of course, there are also
will. But in physics, too, for instance, most models models, which do not show such invariances in their
start discretely: even disregarding the fact that the intended scaling domain. These should not be mod-
universe is finite in principle and structured in parti- elled by calculus. Such situations arise in consid-
cles, and that there are quanta (i.e. smallest unit#s), erations about fractal phenomena: the length of a
it is a fact for quantities which are usually conceived coastline (as a quantity of integral type) is typically
of as being continuous, and mathematized accord- not invariant with the measuring unit, but of course
ingly, that concrete models based, say, on results of the assumed statistical self-similarity also holds only
measurements, will start as discrete models simply in a sensible scale, which certainly does not extend
because continuous functions cannot be obtained as to the microscopic level.
the results of a series of measurements which yield
(d) The transition from models to concrete nu-
only discrete sequences or time series. (This does
merical results cannot be accomplished in general
not hold, of course, for modeling based on theoreti-
without continuous analysis. This is true, for one
cal approaches.)
thing, because of the rounding errors which in-
(c) The continuous character of models using evitably occur in numerical computing, and have
analysis is the result of the intended domain of valid- to be controlled by a more abstract model which
ity.8 Most mathematical models have a specific in- does not include the discretization error. A second,
tended domain of validity, especially a certain scale deeper reason follows from a closer look at the dis-
level, even if this is not explicitly stated. A Ku- crete aspects mentioned in points (a) and (b): It is
clidean line serves as a model for edges of solid bod- the case that the step widths used in (a) and(b) are
ies, e.g. of a shelf, only at a macroscopic scale. If basically independent of each other, as is to be ex-
we look at such an edge through an electron mi- pected from the argument in (c). The density of the
croscope, the edge doesnt look straight any more, values measured in the measuring process is gener-
and on the atomic scale, it loses its one-dimensional ally determined by practical considerations such aa
character too. Therefore, although the edge is well information content and cost. One of the most
modelled by a line, we should not draw conclusions fundamental hypotheses for determining the step
from this model outside its intended domain of va- width is that a diminution of the step width may
lidity. In an analogue sense, calculus models of dis- yield more exact results, but basically not results
crete real phenomena typically are only intended for which differ in principle. The phenomena which are
phenomena at scales where the discreteness doesnt to be observed and/or described are considered to be
invariant with respect to the step width used in the
7 This is properly described in Artigue [1989]. observations provided it is sufficiently small. This
We have taken this argument from Rice [1988] fits in with the assumption that the corresponding
who writes under the sub-heading Verifiable Hy- limits exist. It is only on the basis of this assump-
potheses: Does Mathematics Model Reality? : . . tion that the measuring process can be carried out
we can argue that the real world is inherently dis- in a discrete way chosen by practical considerations.
continous everywhere, its microscopic structure is In this case, however, the phenomena concerned are
either discrete or random or both. In any case, basically invariant with respect to the step width,
the mathematical definition of continuity, deriva- and are thus best described in mathematical models
tion, etc., do not apply because, at some fine scale which do not explicitly contain a step width. The
of examination, the functions are undefined or dis- fact that the step width with which the measured
crete or something intractable. The implication of data were obtained is only of marginal importance
this view is that the concepts of smoothness and be- for the model explains why step widths used, say,
haviors of functions are related to a scale and that to solve numerically the corresponding differential
an adequate mathematical model must take this into equations, will generally be completely independent
account. (p. 37). of the step width used in measurement. Both are
__---_ll_- -II---u-- ~-tlt..l- ..I .-.- --I ^ ._ _-_ ... __I__(_.__ ._.^ ._ l-l_l-.-__-
112 Id uence of Computers and Informatics on Mathematics and Its Teaching
independently determined by practical criteria such time where new offspring could be noticed.
as cost and the precision required. 2.3 Symbolical, numerical and qualitative
This fundamental consideration has been refor- solutions
mulated here for the special case where the results of Linear differential equations, and some others
discrete measurement are used as a starting point. which may be transformed to those, can be solved
It is true, in an analogous way, pointed out in (c), explicitly by closed formulas. From such formu-
for all the other cases in which mathematizing and las one can - at least in principle - answer almost
modeling is done by analysis. any question about the underlying dynamical sys-
This behaviour is of course not valid for all math- tem: asymptotic behaviour, stability and depen-
ematical models in the sciences or other domains. dence on initial values and parameters. But this
But it is in a sense typical for calculus models: If is the exception, not the rule, since most dynamical
this behaviour is not observed in a specific situa- systems arising from model building are essentially
tion, then normally we should really use discrete nonlinear9 and do not admit any closed-form solu-
models, and if we - for technical reasons - neverthe- tions. Numerical solution algorithms on the other
less use some calculus models, we should be aware hand are generally not sensitive to nonlinearity, but
of the improper use and of possible difficulties in in- they share a double experimental character: in most
terpreting results. This may happen for example if cases, the degree to which they approximate the
we try to consider fractal phenomena in nature, true solution can only be estimated, not proven;
such as natural borders (of islands, leaves of trees, and - more seriously - a numerical solution has a
etc.). Here, for example, the application of formulas strict local empirical character. It does not by it-
for the length of a curve does not make much sense. self allow any conclusions about other initial val-
2.2 The context of dynamical systems ues or parameters, which is catastrophic in applica-
tions where such values are only estimated. So they
Dynamical systems (systems of time-indepen-
necessarily need to be complemented by theoretical,
dent explicit first order ordinary differential equa-
usually qualitative considerations about possible be-
tions) appear as rather natural mathematical mod-
haviours of this or a slightly modified dynamical sys-
els for many situations in a variety of disciplines
tem, be it continous or discrete. So this describes
such as the physical, biological or economic sciences.
another complementarity between discrete numeri-
Here typically we have to distinguish between situa-
cal and theoretical methods.
tions where a natural step width exists whose value
influences the phenomena, and situations in which 3. EXPERIMENTS IN USING INFOR-
this is not the case. In both cases, modeling with MATIC TOOLS
(discrete) difference equations is possible and ad- In this section we report on some experiments
equate; but whereas in the former case, the step concerning the use of informatic tools in teach-
width of the difference equation has to be equal to ing basic mathematical courses at the Politecnico
that of the underlying situation, in the latter it may of Torino (Italy), F acuity of Engineering Sciences.
be chosen as a free parameter which suggests that These experiments refer in particular to the courses
the use of differential equations might be more nat- Mathematical Analysis 1 and Mathematical Anal-
ural. ysis 2 given to students of Mechanical Engineering
As an example, consider the logistic growth of in the years 1980 to 1983, using pocket computers.
a (biological) population. If the generations of the This activity was continued in 1984 and 1985, in
population are distinct, as with certain bugs, there the same courses, using such micro computers as the
may be observed oscillations and fluctuations of the Sharp MZ803 and IBM PC. At this second stage, the
population, which are easily modelled and explained experiment was concerned with a restricted number
in the context of a difference equation, but would of students, selected on the basis of a test.
disappear in the transition to the corresponding dif- The experiment was sufficiently successful so
ferential equation (if it were not explicitly mod- that since 1986 all the students of the course (about
elled by including a time lag which would induce 300) have been taught in the computer enhanced
similar fluctuations but would exclude the result- style. At the Politecnico of Torino an introductory
ing equation from what is normally considered a
differential equation in mathematics). But if gen- For an interesting account of nonlinear model
erations are not distinct and population oscillations building see West [1985].
are slow compared to normal reproduction times, lo An exception is the so-called EEE-methods, see
modeling with (logistic) differential equations seems Kaucher/Miranker [1984], but use of these methods
adequate, even if there were only discrete points in is not yet widespread.
computer laboratory is available for students; engi- to find numerical results; this was done in order to
neering students in the first two years have access check the known results of the theory, and also to
to the lab after the completion of a specific course conjecture new results concerning open problems.
which prepares them for meaningful utilization of To be specific, we briefly list the contents of
the available calculating devices and also supplies the exercise sessions concerning dynamical systems
them with adequate knowledge of a programming (Mathematical Analysis 2):
language. l Cauchy problem for first order ordinary dif-
Several instructors of the Engineering Faculty ferential equations; solutions at the microcomputer,
have experimented with the use of the lab as an aid comparing the methods of Euler and Runge-Kutta.
to the basic first two years mathematics courses, and l First order systems of ordinary differential
from the resulting experience two didactic strate- equations, and in particular autonomous systems;
gies have emerged. One was for the students them- visualization of the trajectories in the phase plane.
selves to perform the actual writing of the software. l Second order ordinary differential equations;
The other was to use existing software. It was ob- solutions on the microcomputer of some nonlinear
served that the preparation of software is, even from equations of particular significance in applications,
a mathematical standpoint, an occasion for investi- such as the pendulum and other equations of math-
gation of the topic at hand. However, a certain risk ematical physics.
was noted in the tendency toward interest in the l A numerical approach and simulation on the
computer itself to the detriment of time intended microcomputer of the trajectories for some problems
for dedication to mathematical reflection. which are still open in their qualitative aspects, as
As far as concerns the use of already available for example the mathematical model of the Lorenz
software packages, the possibilities are many. We attractor.
have readily available software written by colleagues Now we present some further details of some of
instructing in analogous courses, that produced by the above, which appear to us particularly signifi-
students in previous courses, and, of course, soft- cant from the didactic point of view. i) The stu-
ware offered by the companies producing calculating dent, knowing the classical analytic theory of lin-
devices. ear equations with constant coefficients, and having
While we refer to Boieri et al.[1984], to Mascar- some basic notions of the stability theory, is invited
ello-Scarafiotti [1987], [1988] and to Mascarello- to solve the equation 5 + I& + t = 0 on a mi-
Scarafiotti-Teppati [1989] for the general aims, the crocomputer and to visualize the trajectories in the
list of the themes and the results obtained, we phase plane (without any direct assistance from the
should like to detail here some of the topics and con- teacher). Figures 1 and 2 show some drawings of
tent, and to add some final comments, as a proof the kind obtained by a student.
of what we asserted in Section 2.
Let us begin by observing that, to carry out, the
experiment in a useful way, it has been necessary to
rely on basic informatic arguments. To this end, in
the main course of lectures, the teacher, after giv-
ing some notions of the theory of formal languages,
then introduced machine-numbers and algorit,hms
for floating-point arithmetic computations. At the
same time, in this first part of the course, some
proofs of classical analysis results were presented in
computational form.
One of the most important experiments con-
cerned the study of dynamical systems using micro-
computers. More specifically, we began in Mathe-
matical Analysis 1 with the study of discrete dynam-
ical systems, which was introduced after the study
of sequences defined by recurrence formulas. As a Figure 1: For the equation i + 2 = 0, z(0) =
natural continuation, in Mathematical Analysis 2 3, i(0) = 0 the E u 1er method converts what should
we considered continuous dynamical systems, giv- be a circle to an outward spiraling curve.
ing a formal expression of the qualitative results.
Finallv. we returned to the use of microcomm1ters
114 influence of Computers and Lnformetics on Mathematics and Its Teaching
= A
Figure 2: The true solution to i + Z + 0.22 =

0, x(0) = 3, i(0) = 0 is an inwards spiraling curve.
The outward spiraling which results with the Euler
method exactly compensates resulting in no spiral-
ing effect.
A discussion with the students followed concern-
ing the validity of the results obtained in this way;
particularly surprising is the second picture, where
closed trajectories appear for k # 0. ii) The stu-
dents solve on the microcomputer the pendulum
equation &+sin z = 0 by the Runge-Kutta method.
Figure 3 shows the drawing obtained by one student.
Figure 4: Plot of a curve approaching the
Lorentz attractor as obtained by a student using the
modified Euler method with x(0) = 0.00001, y(0) =
0.00001 and r(O) = 0.00001.
In Figure 4 there is a picture obtained by a

student (the completion of the program required
a certain informatic ability, due to the complica-
tions arising from the 3-dimensional represent?tion
of the trajectories in (x, y, z)-space). No compari-
son was attempted with known qualitative results
since the existing literature on the subject seems
to be too far advanced for a second year engineer-
Figure 3: Phase portrait of the undamped pen- ing student. However, a comparison was possible
dulum, 3 + sin x = 0 obtained by a student using a with what might be expected from the physical
Runge-Kutta method. phenomenon (such as fluid turbulence phenomena).
We can observe that the picture seems satisfac- What it is very important to emphasize is that at
tory from a numerical point of view. Some qualita- this stage (end of Mathematical Analysis 2) students
tive aspects of the solutions are underlined by the were able to evaluate correctly the results obtained
teacher, as a check of the known results from the from the computer, namely, to take into account
theory. iii) The student is invited to simulate on the discrepancies which may occur between numeri-
the screen the trajectories of the equation of the cal solutions and analytic solutions, keeping in mind
Lorenz attractor: that the final objective is the interpretation of the
physical phenomenon.
dx/dt = --sx + sy
dy/dt = rx - y - zz with s= lO,r=28,b=8/3 4. CONCLUSIONS
dz/dt = -b.z -I- xy. Our considerations have shown that even today
when internally discrete digital computers are used
for hlndling calculus models (so far as applications
are concerned), continuous analysis cannot be dis-
pensed with when describing problems for which et de llnformatique: Seminaire de Didac-
analysis has been classically used. This, however, tique des Mathimatiques et de llnformatique,
need not lead to the conclusion that analysis educa- Grenoble, Ann&e 1988-1989. Universite de
tion at school or universities should go on as before. Grenoble, LSD-IMAG, Institut Fourier.
Our discussion has shown the function of continuous Artigue, M. and Gautheron, V. [1983]: Systkmes
analysis in applications, and teaching must be done I
differentiels, Etude graphique, Paris: CEDIC.
in such a way that this function is fulfilled. This ISBN 2-7124-0722-g.
requires that the transition from the discrete to the
continuous model and vice versa be experienced by Buchberger, B. [1990]: Should Students Learn Inte-
the students and that the respective particular pos- gration Rules? ACM SIGSAM Bulletin, 24,
sibilities and limitations of the model be perceived. 1, 10-17.
To us, it would seem dishonest to try to explain Boieri, P. et al. [1984]: Personal computers in teach-
to the student the importance of analysis for appli- ing basic mathematical courses, SEFI Annual
cations by means of unrealistic and oversimplified Conference: The Impact of Information Tech-
minimum-maximum tasks. Rather, it seems crucial nology on Engineering Education, Erlangen.
to have the student at least begin to assess the use-
Johnson, D. and Lovis, F. (Eds.) [1987]: Infor-
fulness of the various components of the system of
analysis, i.e. concepts, approaches, calculi, trans-
lation schemes in practical applications. This goal
should be attained by appropriate problem solving
in the classroom; and formal explication should play
1987. Amsterdam: North-Holland. ISBN O-
a subordinate part. It remains to be seen how a
444-70325-X.
balance between the individual components can be
achieved. The following aspects, however, should be Kaucher, E. and Miranker, W.L. [1984]: Self-
included in any case: Validating Numerics for Function Space Prob-
a) The teaching of analysis should include the lems, San Diego: Academic Press.
treatment and study of discrete models. This leads Mascarello, M. and Scarafiotti, A.R. [1987]: Speri-
to numerical computations. It does not necessar- mentazione didattica nel Politecnico di Torino:
ily imply explicit teaching of numerical mathemat- Support0 informatica ai corsi die Analisi
ics, but requires including important basic numeri- Matematica nel biennio della Facolta, I, Edu-
cal facts such as propagation of errors. cazione Matematica, VIII, II, 2, 147-151.
b) Building models is an important activity Mascarello, M. and Scarafiotti, A.R. [1988]: Using
which must not be neglectsed in favour of just in- computers in calculus examples-classes for en-
terpreting models. In particular, this means that gineers, ECM/87 Ed ucational Computing in
the techniques of finding suitable functions are as Mathematics, Eds T.F. Banchoff et al, Ams-
important as discussing functions. terdam: North- Holland, 93-97.
c) The role and function of (continuous) calculus
Mascarello, M. and Scarafiotti, A.R. [1988]: Experi-
must be developed in an appropriate way. It cannot
ments in mathematical education at secondary
be used to obtain numerical results, save in excep-
school.and Politecnico of Torino (Italy), ICME
tional cases: it can, however, guide and direct the
6, Theme Group 2 Computers and the teach-
use of numerical methods.
ing of mathematics, Working Group B.5 The
d) The recent development of computer science
effects of technology and of computer science
has established techniques, in part,icular program-
on a maths curriculum for the future, prepara-
ming languages, which permit the precise descrip-
tory papers, Budapest, 88-93.
tion even of complicated processes such as, for in-
stance, the algorithms necessary for symbolic dif- Mascarello, M.; Scarafiotti, A.R.; and Teppati, G.
ferentiation. Mathematics teaching should increas- [1989]: Cultura e insegnamento: Esperienre
ingly make use of these results. significative appoggiate a metodi e strumenti
informatici, Proceedings of the congress /I
REFERENCES Cultura Mathematics e Insegnamento/, Uni-
versita di Firenze, CDO, 253-260.
Artigue, M. [1989]: U ne recherche dinglnierie
didactique sur lenseignement des equations Rice, J.R. [1983]: Numerical .Methods, Software,
differentielles en premier cycle universitaire and Analysis, IMSL Reference Edition, New
I
in Equipe de Didactique des Mathema- tiques York: McGraw Hill.
Rice, J.R. [1988]: Mathematical Aspects of Scien-

tific Software in J.R. Rice (Ed.), Mathematical
Aspects of Scientific Software, Springer, New
York and in The IMA Volumes in Mathematics
and Its Applications, 14, 1 - 39.
Tall, D. [1986]: Building and Testing a Cognitive
Approach to the Calculus Using Interactive
Computer Graphics, Ph.D. Thesis in Mathe-
matics Education, The University of Warwick,
Faculty of Education.
Watanabe, S. [1984]: A n experiment toward a gen-
eral quadrature for second order linear ordi-
nary differential equations by symbolic com-
putation in J. Fich (Ed.), EUROSAM 84, In-
ternational Symposium on Symbolic and Alge-
braic Computation, Cambridge, England, July
9 - 11, 1984, Berlin: Springer-Verlag, 13 - 22.
West, B.J. [1985]: An Essay on the Importance of
Being Nonlinear, Lecture Notes in Biomathe-
matics 62, Berlin: Springer-Verlag. ISBN 3-
540- 16038-B.
Winkelmann, B. [1984]: The Impact of the Com-
puter on the Teaching of Analysis, Int. J.
Math. Educ. Sci. Tech., 15, 675 - 689.
Winkelmann, B. [1989]: Dynamische Systeme und
Differentialgleichungen. Einige mathemati-
sche Anmerkungen zu dynamischen Systemen
und ihrer Simulation, LOG IN 9, Heft 4, 19 -
23. ISSN 0720-8642.
GRAPHIC INSIGHT INTO MATHEMATICAL CONCEPTS
David Tall
Mathematics Education Research Centre
Warwick University, Coventry, U.K.
Beverly West
Mathematics Department
Cornell University, Ithaca, NY 14583, U.S.A.
The human brain is powerfully equipped to pro- Although a graph may be curved, under high mag-
cess visual information. By using computer graph- nification a small part may well look almost straight.
ics it is possible to tap this power to help students In such a case we may speak of the gradient of the
gain a greater understanding of many mathematical graph as being the gradient of this magnified (ap-
concepts. Furthermore, dynamic representations of proximately straight) portion. For instance, a tiny
mathematical processes furnish a degree of psycho- part of the graph y = x2 near x = 1 magnifies to a
logical reality that enables the mind to manipulate line segment of gradient 2 (figure 1).
them in a far more fruitful way than could ever be
yl;.X128
7
2
achieved starting from a static text and pictures in f <x)=x
a book or roughly drawn pictures on a chalk board 1
or overhead projector.
student exploration
Add to this the possibility
using prepared software and the
of
1.81
sum total is a potent new force in the mathematics
curriculum.
A-.:,
0.99
In this paper we report on the development of in-
teractive high resolution graphics approaches to var-
ious areas in mathematics.
centrated initially
The first author has con-
on the calculus in the UK (Tall,
LA!L
X=1
0. 8 1 1.81
y=l
1986, Tall et al, 1990) and the second is working
in the USA on differential equations with John H. L
Hubbard (Hubbard and West, 1990).

Figure 1: Magnifying a small part of a graph to
An interactive visual approach is proving suc- show its local straightness.
cessful in other areas, for example, in geometry (The
Geometric Supposer, Cabri Geomttre), in data ma- To represent the changing gradient of a graph,
nipulation (e.g. Macspin, Mouse Plotter), in prob- it is a simple matter to calculate the expression
ability and statistics (e.g. Robinson and Bowman, (f(x + c> - f(x))/ c f or a small fixed value of c as
1987) and, more generally, in a wide variety of top- c varies. As the chord clicks along the graph for
ics (such as the publications in the Computer Illus- increasing values of z, the numerical value of the
trated Text series, which use computer programs to gradient for each successive chord can be plotted
provide dynamic illustrations of mathematical con- as a point and the points outline the graph of the
cepts). gradient function (figure 2). In this case the chord
gradient function of sin z for small c approximates
New approaches to mathematics to cos 2, which may be checked by superimposing
the graph of the latter for comparison. Thus the
The existence of interactive visual software leads
to the possibility of an exploratory approach to gradient of the graph may be investigated experi-
mentally before any of the traditional formalities of
mathematics which enables the user to gain intuitive
insight into concepts, providing a cognitive founda- limiting processes are introduced.
tion on which meaningful mathematical theories can Such moving graphics also enable the student to
be built. For example, the notion of a limit has tra- get a dynamic idea of the changing gradient. Stu-
ditionally caused students problems (e.g. Cornu, dents following this approach can see the gradient as
1981, Tall and Vinner, 1981). The computer brings a global function, not simply something calculated
new possibilities to the fore; we may begin by con- at each individual point.
sidering the gradient not of the tangent, or of a The symbols dx, dy can also be given a meaning
chord as it approaches a tangential position, but as the increments in 2, y to the tangent. Better
simply the gradient (or slope) of the graph itself. still, (dz,dy) may be viewed u the tangent vector,
117
a valuable idea when we come to the meaning of

differential equations.
f <x)=sinx
from x=-n to ?I
gradient function
~f~r+c~-fCx~~/c
f 01
c=1/1e
Figure 3: Building up the blancmange function
adding successive half-size sawtooth graphs.
The process may be shown dynamically on a vi-
.- L.--
sual display unit; we regret that it cannot be pic-
Figure 2: Building up the gradient function of a tured satisfactorily in a book. But higher magni-
graph. fication of the blancmange function using prepared
software shows it can nowhere be magnified to look
straight, so it is nowhere differentiable. This in-
tuitive approach can easily be transformed into a
Conceptualizing non-differentiable functions
formal proof of disarming simplicity (Tall 1982).
In a traditional calculus course, non-differenti- Visualizing solutions of first order differential

able functions would not be considered until a very equations
late stage, if at all. However, if one views a differen- In graphical terms, a first order differential equa-
tiable function as one which is locally straight, tion dy/dx = f(x,y) simply states the gradient of
then a non-differentiable function is simply one a solution curve at any point (x, y) and a solution
which is not locally straight. For instance, the graph is simply a curve which has the required gradient
of Ix - 11 at z = 1, or Isinxl at x = 7r, has a cor- everywhere. The Solution Sketcher (Tall 1991) or
ner at the point concerned with different, gradients MacMath (Hubbard and West 1991) allows the user
to the left and right. More generally, it is possible to point at any position in the plane and draws
to draw a function that is so wrinkled that it never a small line segment of the appropriate direction.
looks straight anywhere under high magnification. This line-segment may be marked on-screen and
successive line segments fitted together to build up
An example is the blancmange function bl(x),
an approximate solution curve. More broadly, it is
first constructed by Takagi in 1903. First a saw-
possible to draw a direction diagram with an array
tooth S(Z) is constructed for a real number x by
of such segments and to trace a solution by following
taking its decimal part d = L - IX?(x) and defining
the given directions (figure 4).
d ifd< $ dy/dx=-x/y
s(x) =
1- d otherwise. d ,,/, --/-- -...-.-.,\, i mPr*owed
step by step
step
The sequence of functions 8.2
x=-2.6107
r=3.8315
h(x) = 4x1 dy/dx

=a.8612
b2(x) = s(x) + 5(2x)/2

step 0. 188
b,(x) = s(x) + + s(2-x)/2-
Figure 4: Drawing a numerical solution of a first

tends to the blancmange function (figure 3) order differential equation
Graphic Insight into Mathematics 119
The differential equation
has implicit solutions of the form x2+y2 = k, rather

than an explicit global solution of the form y = f(x).
At points where the flow-lines meet the x-axis, the
tangents are vertical and the interpretation of dy/dx
as a function fails, but the vector direction (dx, dy)
is valid with dx = 0 and dy # 0. Thus a first-order
differential equation is sometimes better viewed in
terms of the direction of the tangent to a solution
curve rather than specifying the derivative.
Figure 5: A family of solutions of a differential
Existence of solutions equation, showing funnel and antifunnel behaviour.
There comes a time in every university course

on differential equations when honesty should com-
pel the teacher to admit that cookbook methods for
solving differential equations are inadequate. Such
innocent looking equations as
dy/dx = y2 - x, dy/dx = sin(xy), dy/dx = exy
do not have solutions that can be written in terms

of elementary functions. Students often mistakenly
confuse this with the idea that the equations have no
solutions at all. However, if they are able to inter-
act with a computer program that plots a direction
field and then draws solutions numerically following
the direction lines, the notion of a solution takes on
a genuine meaning: Of course the equations have
solutions: we can see them! From this cognitive
base it is possible to use the computer to analyse
solutions in an entirely new way.
Qualitative analysis of differential equations
New forms of analysis emerge now that we can

see as many solutions as we wish all at the same
time. In figure 5, notice how the solutions tend to
funnel together moving to the lower right-hand
side; in the upper right they spray apart (an an-
tifunnel). Qualitatively descriptive terms such as
funnel and antifunnel can be defined precisely
to give powerful theorems with accurate quantita-
tive results (Hubbard and West 1991). For exam-
ple, the equation dy/dt = y2 - t in figure 5 has two 3
PI I 1 !31Q
overall behaviours: solutions either approach verti-
cal asymptotes for finite t or fall into the funnel and
approach y = -I/? as t -+ +co. In the antifun-
nel there is a unique solution approaching y = +&
(b)
which separates the two usual behaviours. Further- Figure 6: A numerical approximation to the
more, the qualitative techniques enable us to esti- many-body problem. (a) Masses in initial position
mate the vertical asymptote for a solution through with velocity vectors. (b) A little later under the
any given point with any desired precision. action of Newtons Laws.
Newtons Laws scopes. Graphic images are the key. Its

masochism for a mathematician to do with-
The classical three body problem defies elemen-
out pictures one chaos specialist would say.
tary analysis, yet a computer program can cope with
How can they see the relationship between
relative ease. The program Planets (Hubbard and
that motion and this, how can they develop
West 1990) t a k es a configuration of up to ten bodies
intuition?. (Gleick 1987, pp. 38-39)
with specified mass, initial position and velocity and
displays the movement under Newtons laws (figure Systems of differential equations
6). The data can be input either graphically with
the cursor, or numerically in a table. The program The MacMath software of Hubbard and West
allows exploration of possible planetary configura- (1991) draws solutions of systems of differential
tions and it soon becomes plain that stability is the equations dx/dt = f(x,y),dy/dt = g(x,y) in the
exception rather than the rule. One may wonder I, y- plane and also locates singular points using
under what circumstances stability occurs. Other Newtons method, drawing separatrices for saddle
questions arise, such as the reason for the braided points (figure 8).
rings of Saturn that were a great surprise when first
observed by the Voyager space flight. Nobody had
imagined such a behaviour beforehand, yet braided
behaviour showed up in the very first experiments
with the Planets program.
Figure 7 shows a model of a possible orbit, of a
tiny satellite around two larger bodies, alternately
oscillating between revolving round one then mov-
ing into a position of superior gravitational pull of
the other and moving, for a time, to revolve round
the other (Kocak 1986). Once again, computer ex-
ploration shows vividly how three bodies move in a
complex pattern.
The theory of dynamical syst,ems and chaos Figure 8: Locating singular points and separa-
is a paradigmatic example of a new branch of trices for saddle points.
mathematics in which the complementary roles of
In this way the computer may be used to draw
computer-generated experiments to suggest theo-
solutions of systems of differential equations that are
rems and formal mathemat,ical proofs to establish
far too complicated to draw by hand. As a further
them with logical precision go hand in hand.
example, Artigue and Gautheron (1983) draw the
Chaos has become not just a theory but also solutions of the polar differential equations
a method, not just a canon of beliefs but
dr d0
also a way of doing science. To chaos re- -=sinr,-=cos~
dt dt
searchers, mathematics has become an exper-
imental science, with the comput(er replac- which exhibit limit cycles for T = kw (figure 9).
ing laboratories full of test tubes and micro-
Figure 7: A numerical plot represent,ing a tiny Figure 9: Limit cycles of simultaneous polar dif-
satellite orbiting two larger bodies. ferent,ial equations.
Generalizing the concept of visual solutions
dx/dt=u
A second order differential equation such as dufdt =-x
d2x -t
z=
no longer has a simple direction field in (t, x) space,
because through each point, (t, x) there is a differ-
ent solution for each starting direction u = dxjdt.
However, this differential equation is equivalent to
the simultaneous linear equations:
dx
dt=v
dv -t Figure 10: Two nearby solution curves for a pair
z= of simultaneous differential equations.
and in three dimensions, with coordinates (t, t, u),

these equations determine a unique tangent vector
(dt, dx, dv) in th e d irection (1, v, -t). Hence the idea
of a direction field does generalize, but it must be vi-
sualized in three-dimensional (t, x, V) space. Figure
10 shows two solutions of the simultaneous differ-
ential equation spiralling through (t, x, v) space and
their projections onto the t-x and t-u planes, with
the t - z projection giving solutions to the original
second order differential equation.
Visual exploration in geometry
Figure 11: Cabri GCom&tre software for manip-

Euclidean geometry traditionally served to in- ulating geometric figures.
troduce students to a deductive system. In many
countries (such as the United Kingdom) it has all
but disappeared from the mathematics curriculum.
Computers now give the opportunity to manipulate and displayed. Though only represented as a projec-
geometrical figures to build up intuitions for pos- tion of three dimensions onto the two-dimensional
sible theorems (the Geometric Supposer, Schwartz screen, the data may be rotated and viewed dynam-
and Yerushalmy, 1985, Cabri Geomktre, 1987). The ically from any angle to give a sense of depth that is
initial phase of study of geometry can now be an not visible in a static picture (figure 12). Individual
experimental science, in which the student can use points may be selected and inspected to see where
the computer to construct a figure and experiment the data originates to identify interesting informa-
with it. tion, such as outlying values. Rotating the data in
the figure suggests that it clusters together in a way
which intimates that the three components are cor-
Visual Data Processing related.
Modern spreadsheets, statistical packages and

It is now possible to explore data visually, for data handling packages now include visual repre-
example, to see a line of best fit for data in two or sentation of data which encourages the user to ex-
three dimensions. MacSpin allows up to ten cate- plore and communicate complex information in vi-
gories of data, from which any three can be selected sual ways.
and thereby come to understand various mathe-

matical concepts (see, for example, Dubinsky and
Schwingendorf 1991). Clearly a spectrum of ap-
proaches may be possible with varying amounts of
programming, depending on the time and commit-
ment available.
New Styles of Learning

Software is becoming widely available to give
graphical representations in calculus, differential
equations, geometry, data handling, numerical anal-
ysis, and many other areas of mathematics. This is
usually predicated on a new kind of learning ex-
perience & one in which the student may explore
and manipulate ideas, investigate patterns, conjec-
Figure 12: Manipulating data with three com-
tnre theorems and test theories experimentally be-
ponents to look for a visual correlation.
fore going on to prove them in a more formal con-
The ability to present and manipulate informa- text.
tion visually is becoming widely available in many For instance, beginning calculus students may
different areas in mathematics. For example, Robin- investigate the gradients of functions such as sine,
son and Bowman (1986) introduce probability and cosine, tangent, exponential and logarithm, and
statistics using computer graphics with the inten- conjecture their formulas before they are derived
tion of giving a feel for probability distributions formally (Tall 1986, 1987). In differential equations
rather than elaborating mathematical detail. More they may explore problems at the boundaries of re-
generally, the Computer Illustrated Texts (starting search (such as the rings of Saturn) and make the
with Harding 1985) are designed to use simple com- mental link between the friendly world of (mostly
puter programs to provide interactive illustrations linear) equations that can be solved by formulas and
of mathematical ideas which can be explored by the the strange world of those (usually nonlinear) that
student in place of static pictures in a book. can not (Hubbard and West 1991).
This form of learning is not a replacement for
Is programming essential?
formal deduction, but a precursor and a comple-
We have not explicitly mentioned programming ment to it. It enables the less able student to grasp
for the purpose of gaining insight into mathematical essential ideas that would previously be too diffi-
processes. A body of expertise is growing in which cult when framed in a purely formal theory and for
students are expected to write or adapt short pro- the more able student to build a cognitive base for
grams (usually in structured Basic, Pascal, or Logo) the formal theory to follow. It enables a wide range
to carry out mathematical algorithms. From here it of students to integrate their knowledge structure
is often intended that they move on to prepared soft- through their powers of visualization.
ware that uses the underlying algorithms in a more
interactive manner. The early computer-illustrated Acknowledgement
texts assumed that the programming would be suf- The authors are grateful to Professor John H.
ficiently simple that it would allow the student to Hubbard for his assistance in the preparation of an
modify the programs, but this became an impossible earlier version of this article.
ideal in later texts as more sophisticated programs
were written that were too complex for the user to REFERENCES
modify. Programming requires a serious investment Alfors, D. and West, B. [1992]: Analyser* calculus
in time and effort. However, it can pay vast divi- software (for the Macintosh computer), Read-
dends in gaining insight into the underlying math- ing, MA: Addison-Wesley.
ematical processes if the investment is sufficiently
Artigue, M. and G?utheron, V. [1983]: Systtmes
generous.
Diffirentiels: Etude Graphique, CEDIC, Paris.
Dubinsky has evidence that having students
make certain programming constructions (in the Bach, J.O. [1990]: MultiMat (for I.B.M. compatible
computer language ISETL) can lead to their making computers; available in English), Copenhagen:
parallel mathematical constructions in their minds Danish Mathematics Teachers Assbciation.
Cabri GComktre [1987]: IMAG, BP 53X, Universite Schwartz, J. and Yerushalmy, M. [1985]: The Ge-
de Grenoble (for IBM and Macintosh comput- ometric Supposer, Sunburst Communications,
ers). Pleasantville, N.Y. (for the Apple Computer).
Cornu, B. [1981]: Apprentissage de la notion de lim- Takagi, T. [1903]: A simple example of a continu-
ite : modkles spontanes et modeles propres, ous function without derivative, Proc. Phys.-
Actes du Cinquikme Colloque du Groupe In- Math., Japan, 1, 176-177.
ternational P.M. E., Grenoble, 322-326. Tall, D.O. [1982]: The bl ancmange function, con-
Devaney, R.L. [1990]: Chaos, Fractals, and Dynam- tinuous everywhere but differentiable nowhere,
ics: Computer Experiments in Mathematics, Mathematical Gazette, 66, 11-22.
Reading, MA: Addison-Wesley. Tall, D.O. [1986]: Graphic Calculus I-III (for BBC
Dubinsky, E. and Schwingendorf, K. E. [1991]: Con- compatible computers), Glentop Press, Lon-
structing calculus concepts: cooperation in a don.
computer laboratory in The Laboratory Ap- Tall, D.O. [1987]: R ea dangs in Mathematical Educa-
proach to Teaching Calculus (Leinbach, L.C., tion: Understanding the Calculus, (collected
Ed.), MAA Notes Series No. 20, Washington, articles from Mathematics Teaching, 1985-7),
DC: Mathematical Association of America. Association of Teachers of Mathematics, UK.
Gleick, J. [1987]: Chaos: Making a New Science, Tall, D.O. [1991]: Real Functions and Graphs
London: Penguin and Viking: New York. (for BBC compatible computers), Cambridge:
Harding, R. [1985]: Fourier Series and ?Fans- Cambridge University Press.
forms, A Computer Illustrated Text, Bristol Tall, D.O., Blokland, P. and Kok, D. [1990]: A
and Boston: Adam Hilger (for the BBC, JBM Graphic Approach to the Calculus (for I.B.M.
and Apple Computers). compatible computers), Sunburst Inc., USA.
Hubbard, J.H. and Parmet, M. [1987]: 3D Analyzer; Tall, D.O. and Vinner, S. [1981]: Concept image
Complex Paint (for the Macintosh computer), and concept definition in mathematics, with
Ithaca, NY: Cornell University. special reference to limits and continuity, Ed-
ucational Studies in Mathematics, 12 151-169.
Hubbard, J.H. and West, B. [1991]: Diflerential
Equations: A Dynamical Systems Approach, Zimmerman, W. and Cunningham, D. [1990]: Visu-
New York: Springer-Verlag. alization in Teaching and Learning Mathemat-
ics, MAA Notes Number 19, Washington,DC:
Hubbard, J.H. and West, B. [1991]: MacMath: A
Dynamics Package (for the Macintosh com-
puter), New York: Springer-Verlag.
Johnson, J.A. [1990]: Gyrographics (for I.B.M. com-
patible computers), Stillwater, OK: Cipher
Systems.
Klotz, E. and Jackiw, N. [1991]: The Geome-
ters Sketchpad (for the Macintosh computer),
Berkeley, CA: Key Curriculum Press.
KoCak, H. [1986]: Phaser: Diflerential and DiJ
ference Equations through Computer Exper-
iments (for IBM computers), New York:
Springer-Verlag.
Ma&pin [1985]: D2 Software Inc., Austin, TX (for
the Macintosh Computer).
Phillips, R. [1988]: Mouse Plotter, Shell Centre,
Nottingham (for the Archimedes computer).
Robinson, D.A. and Bowman, A.W. [1986]: In-
troduction to Probability, Bristol and Boston:
Adam Hilger (for the BBC and IBM comput-
ers).
Annotated References
This is a list of some of the references earlier in this book of particular significance or usefulness
with a brief annotation describing the contents of the article or book.
Aho, A.V., Hopcroft, J.E. and Ullman, J.D. [1983]: Considers a wide range of symbolic compu-
Data Structures and Algorithms, Reading, tation activities with examples including com-
MA: Addison-Wesley. puter algebra, computational geometry, auto-
matic reasoning and automatic programming.
This is a graduate text that presents the
data structures and algorithms that underpin
much of todays computer programming. It Banchoff, T. et al (Eds.) [1988]: Educational Com-
covers these topics in the context of solving puting in Mathematics, ECM87, Amsterdam:
problems using computers and introduces step North Holland.
counting and complexity as an integral part of
problem solving. It is very nicely written and Proceedings of an international congress
comprehensive. with presentations about a number of math-
ematical domains (differential geometry, cal-
culus, dynamical systems, geometry, etc.) and
Alfors, D. and West, B. [1992]: Analyzer* calculus
reports about various teaching experiments.
software (for the Macintosh computer), Read-
ing, MA: Addison- Wesley.
Bolter, J.D. [1984]: Turings Man: Western Cul-
EDUCOM/NCRIPTAL distinguished soft-
ture in the Computer Age, Chapel Hill, NC:
ware for 1990 - for calculus and iteration of
University of North Carolina Press.
functions of a single variable.
An innovative analysis of the philosophical
Artigue, M. [1989]: U ne recherche dinginierie impact of computers as embodied mathemat-
didactique sur lenseignement des Cquations ics, of nature as information, and of humans
differentielles en premier cycle universitaire in a23 information processors. Bolter draws
kquipe de Didactique des Mathematiques et de interesting parallels with Athenian man-
llnformatique: Siminaire de Didactique des limited by finite mathematics, yet creating
Mathimatiques et de llnformatique, Greno- private universes like the craftsman potter in
ble, An&e 1988-1989. Universitd de Grenoble, Platos Timaeus.
LSD-IMAG, Institut Fourier.
A didactical reflection on the prevalence Bush&w, D. [1983]: A two-year lower-division math-

of the algebraic approach to differential equa- ematics sequence in The Future of College
tions in undergraduate courses and construc- Mathematics, A. Ralston and G.S. Young
tion of a new course with emphasis on geomet- (Eds.), pp 111-118. New York: Springer-
ric, qualitat(ive and numerical elements. Verlag
An article which presents an outline of

Artigue, M. and Gautheron, V. [1983]: Systemes
a two-year undergraduate sequence which in-
Diflerentielles: Etude Graphique, CEDEC,
tegrates calculus and discrete mathematics.
Paris.
Twenty-four different modules are listed as
A pioneer work that deserves to be a clas- well as the interrelations among them.
sic. Teaches qualitative methods in the study
of two-dimensional systems of autonomous dif- Clocksin, W.F. and C.S. Mellish [1981]: Program-
ferential equations with many beautiful de- ming in Prolog, Berlin: Springer-Verlag.
signs.
This is a textbook for teaching Prolog as
Aspetsberger, K. and Kutzler, B. [1988]: Symbolic a programming language. It briefly discusses
com.putation - A new chance for education the logical foundations of the language but
in F. Lovis and E.D. Tagg (eds) Computers the emphasis is on how useful programs can
in Education., 331-336, Amsterdam: North- be written using the Prolog systems that are
Holland. available.
124
Computers and Mathematics, A column (past edi- Dubinsky, E. and Schwingendorf, K. E. [1991]: Con-
tor: J. Barwise, current editor: K. Devlin) ap- structing calculus concepts: cooperation in a
pearing regularly in the Notices of the Ameri- computer laboratory in The Laboratory Ap-
can Mathematical Society. proach to Teaching Calculus, (Leinbach, L.C.,
Ed.), MAA Notes Series, Washington, DC:
Papers appearing in this column discuss all Mathematical Association of America.
aspects of the influence of computers on math-
ematics, both with respect to research and to Describes an early version of the Purdue
University calculus project which emphasizes
teaching. Many Symbolic Mathematical Sys-
cooperative learning. Contains examples of
tems have been discussed - see Appendix 1 of
the paper by B.R. Hodgson and E.R. Muller computer assignments and examinations as
in this volume. well as details of the computer laboratory set-
up*
Cornu, l3. [1992]: LOrdinateur pour enseigner les Fey, J.T. [1989]: Technology and Mathematics Edu-
m@hk,matiques, Paris: Presses Universitaires cation: A Survey of Recent Developments and
de France. Important Problems, Educ. Studies in Math.,
20, 237-272.
A multi-author book with chapters on
Provides an overview and analysis of re-
mathematics and informatics, on various ex-
cent progress in applying electronic informa-
amples of teaching mathematical concepts,
tion technology to the creation of new envi-
and on the link between technology and ed-
ronments for intellectual work in mathematics.
ucational research.
It discusses numerical computation, graphic
computation and symbolic computation and
Devaney, R.L. [1990]: Chaos, Fractals, and Dynam- contains many concrete examples.
ics: Computer Experiments in Mathematics,
Reading, MA: Addison-Wesley. Foster, K.R. and Bau, H. H. [1989]: Symbolic Ma-
nipulation Programs for the Persongl Com-
Paperback handbook full of explorations puter, Science, 243, 679-684.
for either a classroom setting or individual
Provides a useful comparative summary
study.
of the capabilities of a majority of Symbolic
Mathematical Systems. The article provides
Douglas, R. (Ed.) [1986]: Toward a Lean and prices and types of hardware required. (These
Lively Calculus, Proceedings of a Confer- systems are constantly changing and up to
ence/Workshop at Tulane University, Wash- date information should be obtained from the
ington, DC: Mathematical Association of manufacturers whose addresses are listed in
America. the article.)
The report of a conference - sometimes Garey, M.R. and D.S. Johnson [1978]: Computers
called the counterreformation (in relation to and Intractability, A Guide to the Theory of
discrete mathematics) - in which various NP-Completeness, San Francisco: Freeman.
recipes for improving the teaching of calculus
This book includes a thorough introduc-
are given.
tion to complexity theory and AfP-complete
problems. It shows how to recognize these
Dubinsky, E. and Fraser, R. [1990]: Computers problems and how to deal with them. It is
and the Teaching of Mathematics, Notting- a readable guide and offers an exhaustive list
ham, UK: The Shell Centre for Mathematical of NP-complete problems.
Education.
Goldenberg, E.P. [1988]: Mathematics, Metaphors,
This selection of papers from the Sixth In- and Human Factors: Mathematical, Techni-
ternational Congress on Mathematics Educa- cal, and Pedagogical Challenges in the Edu-
tion contains those papers from ICMEG par- cational Use of Graphical Representation of
ticularly relevant to the use of technology in Functions, Journal of Mathematical Behavior,
teaching mathematics. 7, 2, 135-173.
______ -__-..-__ .-._-. .--.

Describes the work of a research group at Twelve interactive and easy-to-use graph-
the Educational Technology Center (ETC) at ics programs for both iteration and differen-
the Harvard Graduate School of Education. tial equations, with a handbook of sugges-
Many convincing examples show possible pit- tions for what to do with them; developed
falls and misunderstandings in the interpreta- to accompany the authors three volume text
tion of function graphs. Differential Equations: A Dynamical Systems
Approachalso being published by Springer-
Verlag.
Gray, T.W. and Glynn, J. [1991]: Exploring Math-
ematics with Mathematics, Reading, MA: Jacobsen, E. [1989]: An International Perspective in
Addison-Wesley. SIGCUE OUTLOOK (Bulletin of the Special
Interest Group for Computer Uses in Educa-
An introduction to Mathematics, written
tion), 20, 2, New York: ACM Press.
in the form of a dialog between the two au-
thors, presenting exploration of various math- Provides a worldwide overview of micro-
ematical concepts. A section is devoted computers in education. The priorities of gov-
to Mathematicss application to high school, ernments and current usages are illustrated by
college and university mathematics teaching. selecting representative countries in each con-
The book comes with a CD-ROM disk con- tinent. UNESCOs role in the field of comput-
taining an electronic edition of the text in the ers and education is examined.
form of Mathematicss Notebooks.
Jaffe, A. [1984]: Ordering the Universe: The Role of
Mathematics in Renewing U.S. Mathematics,
Heid, M.K., Sheets C. and Matras, M.A. [1990]: Washington, DC: National Academy Press.
Computer-enhanced Algebra: New roles and
challenges for teachers and students in T.J. A sweeping survey of contemporary math-
Cooney and C.R. Hirsch (eds.) Teaching ematics, an ancient art, . . highly esoteric,
and Learning Mathematics in the 199Os, (1990 and the most intensely practical of human en-
NCTM Yearbook), 194-204, Reston, VA: Na- deavors. Emphasizes the role of mathemat-
tional Council of Teachers of Mathematics. ics in advancing computation, communication,
physics, and engineering.
Discusses the new roles for teachers and
students when the computer enters actively Johnson, D.C. and Lovis, F. (Eds.) [1987]: Infor-
into mathematics education. The teacher as matics and the Teaching of Mathematics, Pro-
technical assistant. The teacher as collabora- ceedings of the IFIP TC 3/WG 3.1 Working
tor. The teacher as facilitator and catalyst . Conference on Informatics and the Teaching
Responsibilities for evaluating student learn- of Mathematics, Sofia, Bulgaria, 16 - 18 May,
ing, etc. 1987. Amsterdam: North-Holland. ISBN O-
444-70325-X.
Hirst, A. and Hirst, K. (Eds.) [1988]: Proceedings Many international contributions on a wide
of the Sixth International Congress on Mathe- scope of themes, from theoretical to concrete,
matics Education, Budapest, 1988, Budapest: from primary school to college level.
Janos Bolyai Mathematical Society.
Kenney, M., (Ed.), [1991]: Discrete Mathematics
The state-of-the-art on the use of com- across the Curriculum, K-12 (1991 NCTM
puters in mathematics education. Contains Yearbook), Reston, VA: National Council of
reports of plenary sessions on computerization Teachers of Mathematics.
of schools and mathematics education, and on
Discrete mathematics is the part of math-
algoritmic mathematics, and of a theme group
ematics where algorithmics are most natural.
on comput,ers and the teaching of mathemat-
This yearbook contains articles are specific im-
ics.
plementation ideas at various levels.
Hubbard, J.H. and West, B. [1991]: MacMath: A Klotz, E. and Jackiw, N. [1991]: The Geome-
Dynamics Package (for the Macintosh com- ters Sketchpad (for the Macintosh computer),
put,er), New York: Springer-Verlag. Berkeley, CA: Key Curriculum Press.
Developed at Swarthmore College as part This issue on the use of function plotters
of the Visual Geometry Project under the di- in mathematics teaching at schools gives gen-
rection of E. Klotz and D. Schattschneider. A eral considerations on the didactical criteria,
dynamic tool for exploring geometry (e.g., you possibilities and limits of such tools, some con-
can construct a figure, distort it, and your concrete examples for the use at various places in
struction follows the distortion.) the mathematical curriculum, and critical de-
scriptions of some function plotters available
Knuth, D.E. [1973]: The Art of Computer Program- in Germany.
ming: vol 9: Sorting and Searching, Reading,
MA: Addison-Wesley. Malkevitch, J. et al., [1988]: For All Practical Pur-
This is the third of Knuths classical books. pose, San Francisco: W. H. Freeman. Also
As with the other two, it should be consid- available as 26 videotaped TV programs.
ered as a source for information and references
An innovative general studies secondary-
and not as a basic textbook. Although the
tertiary course which shows students many
title may sound as if the book is mainly for
current applications of elementary mathemat-
programmers concerned with preparing sort-
ics and shows faculty how many new topics
ing routines, it virtually covers all theoretical
(often algorithmic) can be included in the cur-
aspects of programming.
riculum.
Kocak, H. [1986]: Phaser: Diflerential and Dif-
ference Equations through Computer Exper- Mathematical Sciences Education Board [1990]: Re-
iments (for IBM computers), New York: shaping School Mathematics: A Philosophy
Springer-Verlag. and Framework for Curriculum, Washington,
DC: National Academy Press.
The classic entry into the field of experi-
menting with differential equations. A detailed rationa1.e for changing school
mathematics, building on research related to
Koerner, J.D. (Ed.) [1981]: The New Liberal Arts: the role of technology and to the prpcess of
An Exchange of Views, New York: Alfred P. teaching and learning. Summarizes (with ex-
Sloan Foundation. tensive references) the relevant research litera-
ture; poses open questions; and outlines goals
A position paper followed by ten responses
for curriculum reform.
in which it is argued that analytic skills (e.g.,
statistics, computation, applied mathematics)
are as crucial for liberal education as are tradi- Maurer, S. [1984]: T wo meanings of algorithmic
tional literary, historical, and artistic studies. mathematics, Math. Teacher, 77, 430-435.
The computer has altered the world in which
Explains at length the difference between
the student will live as well as the manner in
the traditional and contemporary meanings of
which he will think about the world.
algorithmic mathematics. The two main ex-
amples are polynomial evaluation (mentioned
Krivine, J.L. and M. Parigot [1990]: Programming
briefly in Maurers article here) and Gaussian
with proofs, J. Inf. Process. Cybern. EIK 26,
elimination for solving systems of linear equa-
149-167.
tions.
This is one of the recent articles which de-
velop the idea that proofs and programs are Maurer, S. [1985] The algorithmic way of life is best,
basically the same object. Even though it is College Math. J., 16, 2-18 (Forum article and
mainly written for readers with a strong back- reply to responses).
ground in logic and computer science, it has
deep insight on the subject. The author presents a deliberately force-
ful argument for abandoning the traditional
LSW [1990]: L an d esinstitut fiir Schule und Weiter- theory/computation schism in favor of an al-
bildung (LSW), Soest (Hrsg.): Neue Medien gorithmics synthesis. The article is followed
im Unterricht - Funktionenplotter im Math- by numerous thoughtful responses, some quite
ematikunterricht. Soest: Soester Verlagskon- critical, and then a summary reply by the au-
tor. ISBN 3-8165-1732-3. thor.
Maurer, S. B. and Ralston, A. [1991]: Discrete Algo- A call for action issued by the National
rithmic Mathematics, Reading, MA: Addison- Academy of Sciences to improve mathemat-
Wesley. ics education in the United States. Highlights
human resource needs, learning through in-
One of the most recent discrete mathemat- volvement, and curriculum priorities. Stresses,
ics texts. It is intended for well-prepared first- among other things, the way computers have
year university students and stresses the algo- changed priorities for mathematics education.
rithmic approach to discrete mathematics.
Mines, B., Richman, F. and Ruitenberg, W. [1988]: lege Education: the HP-28C, Amer. Math.
A Course in Constructive Algebra, New York: Monthly, 94, 895-902.
Springer-Verlag.
Details the power of the HP-28C by pro-
An example of what classical pure mathe- viding examples of its capabilities. Con-
matics may look like if researchers take the al- cludes that it introduces one new element
gorithmic viewpoint to heart. Topics such as into the teaching of mathematics, namely aw-
factorization in polynomial rings are treated some computing power at both modest price
by defining algorithms and proving them cor- and size. (Which is even more true of the
rect. The review by Beeson (see references more recent HP-48SX.)
in Maurer, this volume) gives a thorough
overview. Okamori, H. [1989]: Mathematics Education and
Personal Computers, Tokyo: Daiichi-Hoki
Shuppan.
Muller, E.R. [1991]: Maple Laboratory in a Ser-
vice Calculus Course in L.C. Leinbach et al This volume is an excellent survey of com-
(eds.) The Laboratory Approach to Teaching puter use in mathematics education in Japan
Calculus, MAA Notes Number 20, 111-117, from kindergarten to university. It covers re-
Washington, DC: Mathematical Association of search and practice and includes a number of
America. examples of problem solving in the real world
(e.g. mathematics of a lake, road mathemat-
Presents the development and implementa- its).
tion of compulsory laboratories. Includes ex-
amples of laboratory activities and provides Page, W. [1990]: Computer Algebra Systems: Issues
data on traditional indicators (failure rates, and Inquiries, Computers Math. Applic., 19,
etc.) and student attitudes. 51-69.
An educational-philosophical survey arti-
National Council of Teachers of Mathematics [1989]:
cle on issues of special importance to all who
Curriculum and Evaluation Standards for
are involved with the instructional uses of
School Mathematics, Reston, VA: National
computers in the mathematical sciences.
Council of Teachers of Mathematics.
A key document in Americas attempt to Peressini, A. et al [1992]: Precalculus and Discrete

come from way behind in mathematics educa- Mathematics, University of Chicago School
tion. Contains a detailed set of standards for Mathematics Project, Glenview, IL: Scott,
school mathematics, arranged in four groups Foresman.
(K-4, 5-8, g-12) g iving expectations and ex- The Chicago project has developed a high-
amples in each curricular area. Builds on ly innovative mathematics curriculum for av-
assumption of educating students for an in- erage students in grades 7-12. This book is
formation society; advocates extensive use of the 12th year text. It includes algorithmics.
calculators and computers t,hroughout school There are several other innovative projects un-
mathematics. der way in America which will also result in
texts.
National Research Council [1989]: Everybody
Counts: A Report to the Nation on the Fu- R.alston, A. [1981]: Computer Science, Mathemat-
ture of Matkematics Education, Washington, ics, and the Undergraduate Curricula in Both,
DC: National Academy Press. Amer. Math. Monthly, 88, 472-485.
An urgent appeal to mathematicians to Steen, L.A. (Ed.) [1988]: Calculus for a New Cen-
recognize the fundamental mathematical re- tury, MAA Notes No. 8, Washington, DC:
quirements of computer science by giving dis- Mathematical Association of America.
crete mathematics greater priority in early
A report of a conference whose papers look
years of mathematical preparation. The be-
at the role of calculus and the teaching of cal-
ginning of a decade-long effort to establish dis-
culus as we approach the 21st century.
crete mathematics on an equal footing with
calculus as an important foundation not only
Stern, J. [1990]: Fondements Mathematiques de
for computing but also for mathematics itself.
llnformatique, Paris: McGraw-Hill.
Ralston, A. and Young, G.S. (Eds.) [1983]: The This is an undergraduate textbook which
Future of College Mathematics, New York: covers computability, complexity, logic and the
Springer-Verlag. theory of regular and algebraic sets. It is a
readable introduction to the main tools and
A report of the first conference at which concepts of theoretical computer science.
discrete mathematics as a possible alternative
to or coequal with calculus was considered. Tall, D.O. [1986]: Building and Testing a Cognitive
The papers discuss a wide variety of the rele- Approach to the Calculus Using Interactive
vant issues. Computer Graphics, Ph.D. Thesis in Mathe-
matics Education, The University of Warwick,
Rice, J.R. [1988]: Math ematical Aspects of Scien- Faculty of Education.
tific Software in J.R. Rice (Ed.): Mathemati-
Combining mathematical, psychological
cal Aspects of Scientific Software, New York:
and epistemological studies with the develop-
Springer-Verlag and in The IMA Volumes in
ment of suitable software, important insights
Mathematics and Its Applications, 14, 1 - 39.
into the nature of the learning of the calcu-
Fundamental but concrete aspects of math- lus are gained. The work is not application
ematics, applications and the new role these oriented, but the attempt to build up a true
are taking when the users rely on ready-made understanding using discrete and continuous
mathematical methods. aspects serves the user of mathematics, too.
Tall, D.O. [1986, 19901: Graphic Calculus I-III (for

Robinson, J.A. [1965]: A machine-oriented logic
BBC compatible computers), London: Glen-
based on the resolution principle, J. ACM 12,
top Press and, with P. Blokland and D. Kok,
23-41.
A Geometric Approach to the Calculus (for
This book gives an account of the impres- I.B.M. compatible computers), Sunburst, Inc.,
sive breakthrough achieved by its author to- USA.
wards performing deductive reasoning by a Interactive, extremely well designed, and
machine. It includes the presentation of the
easy-to-use graphics programs for all sorts of
formalism of predicate logic, a thorough expo
calculus topics, including multivariable, with
sition of the resolution principle as well as a excellent manuals for students and teachers.
detailed account of a working computer pro-
gram for showing what follows from what.
Tall, D.O. [1987]: Readings in Mathematical Edu-
cation: Understanding the Calculus, collected
Schmidt, Giinter (Ed.) [1988]: Computer im Math- articles from Mathematics Teaching, 1985-
ematikunterricht, Der Mathematikunterricht 1987, Association of Teachers of Mathematics,
34, Heft 4. ISBN 3-617-24022-4, 19-42. UK.
A special issue of a German journal aimed These are among the earliest writings on
mainly at high school teachers. This issue con- the subject worldwide; they include many ex-
tains three articles discussing the impact of cellent insights and suggestions.
computers on mathematics learning, analysing
software for mathematics teaching, and de- Tinsley, J.D. and van Weert, T.J. (Eds.) [1989]:
scribing a possible use of recursion in teaching Educational Software at the Secondary Level,
calculus. Amsterdam: Elsevier.
130 Influence of Computers and Informatics on lllathematics and Its Teaching
Proceedings of a 1989 IFIP working confer-

ence with many examples of educational soft-
ware and discussions of the trends of software
development and evolution.
Wagon, S. [1991]: Mathemafica in Action, San Fran-

cisco: Freeman.
An example-based introduction to tech-
niques, both elementary and advanced, of us-
ing Mathematics for mathematical computa-
tion and exploration. An underlying theme
of this book is that a computational way of
looking at a mathematical problem or result
yields many benefits.
West, B.J. [1985]: An Essay on the Importance of

Being Nonlinear, Lecture Notes in Biomathe-
matics 62, Berlin: Springer-Verlag. ISBN 3-
540-16038-8.
Fundamental aspects of nonlinearity, most-
ly in the context of dynamical systems. A bit
technical in some parts.
Wilf, H.S. [1982]: The Disk with the College Edu-

cation, Amer. Math. Monthly, 89, 4-8.
An early effort to alert mathematicians to
the power of symbolic computing systems-
which are now much more powerful than those
of a decade ago-and to the threat they pose
for those who might continue the status quo
in teaching undergraduate mathematics.
Zimmerman, W. and Cunningham, D. [1990]: Visu-

alization in Teaching and Learning Mathemat-
Zcs, MAA Notes Number 19, Washington,DC:
An authoritative collection from many ex-
perts; topics include geometry, calculus, dif-
ferential equations, differential geometry, com-
plex analysis, linear algebra, iteration and
stochastic processes.
Zorn, P. [1987]: Computing in Undergraduate

Mathematics, Notices of the American Math-
ematical Society, 34, 917-923 (October).
A careful analysis of issues-philosophical,
pedagogical, practical-associated wit,h the
introduction of computers and computer
labs into undergraduate mathematics courses.
Based on issues raised at a workshop of project
leaders who are seeking to make these changes
on different campuses.
Index
Abstraction, 35 CUPM, 34
Algorithm, 16, 20, 39f, 52f, 70-72, 73 Curriculum, mathematics, 75-76
proof by, 17 in Japan, 76-77
Algorithm analysis, 46-47 in the USA, 77
Algorithm design, 45 primary, 59f
Algorithm verification, 45 secondary, 65f
Algorithmic language, 49 university, 20f, 82f
Algorithmics, 45f Curriculum change, 99-100
Analysis of Algorithms, 46-47
Approximation, 84 Data analysis, exploratory, 15
Arabic multiplication, 39 software for, 70-71
Artificial intelligence, 28 Data structures, 53
Assessment, 6 Databases, 28
AUTOCALC, 63-64 Difference equations, 49
Differential equations, 74, 118
Black boxes, 72-73 systems of, 120-121
Blancmange function, 118 Discovery in mathematics, 23
Bottom-up design, 45 Discovery learning, 5
Bourbaki, 3, 33 Discrete-continuous interplay, 110f
Discrete functions, 82
Calculator exercise, 44
Discrete mathematics, 21, 34, 49, 80, 81f
Calculators for young children, 59
Dynabook, 3
Calculus, 22f, 34, 80
Dynamical systems, 112
computers in, 23
CAYLEY, 13 Early science software, 60-62
Children, calculators for, 59 Equations, numerical solution of, 43
computers for, 59 Errors, 82
Classroom, computers in, 23, 35-36 student, 29
Clique problem, 54 Estimation, 82
Complexity theory, 52 Euclids algorithm, 40, 45, 47
Computability, 51 Eulers method, 113
Computation theory, 51 Experimentation in mathematics, 14, 65-66
Computer-aided design, 58 Exploration in mathematics, 23
Computer-aided instruction, 58 Exploratory data analysis, 15
Computer-aided learning, 110
Computer algebra systems, Fast Fourier transform, 53
see Symbolic mathematical systems Four colour theorem, 14
Computer-assisted learning, 29 Fractal, 16
Computer graphics, 23, 28, 117f Function, 12, 82
Computer literacy, 34-35 blancmange, 118
Computer science, 34, 37 discrete, 82
theoretical, 49, 50 non-differentiable, 118
Computers plotting, 74
experimental use of, 112-l 14 recursive, 52
for mathematics teaching, 25f software, 69-70
in the classroom, 23 Takagi, 13
instruction with, 110
Continuous functions, 82 Geometry, software for, 66-67
Courseware, 29 visual exploration in, 121
131
132 Influence of Computers and Informatics on hlathematics and Its Teaching
(iraplt a.lgorithins. 53 New mathemat,ics, 3

Graph t.lreory, 49 curricula for, 6, 31
Guessing, 36 in Japan, 76-77
in the USA, 77
Hamiltonian path problem, 53 Newtons Laws, 120
Heuristics, 47 Non-differentiable functions, 118
Hypermedia, 28 NP-complete problems, 53
Numerical analysis, 20
Induction, 46, 49 Numerical solution of equations, 43
Instruction, individualized, 28
Int,egration, 82 Operations research, 49
Iterat,ive met,hods, I.5
Partial correctness, 54
Jacobi symbol, 53 Pattern matching, 53
Journey in Mathematics project, 7 Plotting functions, 74
Julia set,, 16 Pocket calculator, 29
Polynomial evaluation, 44, 46
Primality tests, 53
KALEIDOSCOPE, 67-68
Primary schools:
Key handling programs, 62
curriculum, 59f
Kleenes theorem, 52
software for, 60f
Problem solving, 5
Language development software, 60
Program correctness, 54
Laptop computers, 36
Programming, 122
Learning, new styles of, 122
Prolog, 54
Levels of research and development, 9
Proof, 13
Logic, 20, 54f
by algorithm, I7
Loop invariant, 45
Quadratic formula, 42-43

MacMath, 118
Mathematician, student ns, 2 Random numbers, 43
Mathematical communication, 17 Recursion theory, 52
Mathematical induction, 46, 49 Recursive function, 52
Mathematical logic, 20, 54f Representation, 84
Mathematical metaphors, 36-37 Resolution, 54
Mat(hematica1 modelling, I5 Runge-Kutta method, 114
Mathematics, discovery in, 23
experimentation in, 14, 65-66 Satisfiability problem, 53
explorat,ion in, 23 Searching, 49, 53
new approaches to, 117-118 Secondary school curriculum, 65f
science of, 19 Self-evaluation, 28
university, 19 Simplex method, 53
Mathematics programs, 63 Simulation, 15, 73-74
Mathematics teaching, computers for, 25f stochastic, 74
Matrix mult,iplication, 41 Smalltalk, 3
Model building, 74-76 Software, 29f
Models, 111-112 data analysis, 70-71
Moving t,arget problem, 1 for primary schools, 60f
Mult,imedia, 28 for secondary schools, 66f
Index 133
function, 69-70
Soliton, 13
Sorting, 49, 53
Spreadsheet, 28
Square-root construction, 41-42, 47
Stochastic simulation, 74
Student errors, 29
Student-teacher relation, 26
Symbolic mathematical systems, 17, 72, 93f
classroom sessions with, 97
curriculum implications of, 99
examples of use, 94-95, 105-106
individual access to, 98
laboratory sessions with, 97
soft.ware reviews of, 104-105
Symbolic solutions, 112
Syntax analysis, 55
SYMMETRIC TURTLES. 66-67
Takagi function, 13
Teacher, role of, 27
Teacher-student relation, 26
Teacher t,raining, 89f
Teaching, evolution of, 87
Teaching assistant, macro as, 7
Teaching-learning process, 109
Theoretical computer science, 49, 50
Top-down design, 45
Total correctness, 55
Towers of Hanoi, 42, 45, 46, 49
Training methodology, 90-91
Traveling salesman problem, 54
Turing machine, 52, 53
Unification-resolution, 54
University mathematics, curriculum for, 20
preparat#ion for, 19
Verification, algorithm, 45
Visual data processing, 121
Visual exploration, 121
White box, 73
Young children, calculat,ors for, 59

computers for, 59

Eo

Hochgeladen von

Dokumentinformationen

Originaltitel

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

Eo

Hochgeladen von

Copyright:

Verfügbare Formate

Science and technology education, 44

The influence of computers

Published in 1992 by the United Nations Educational,

In 1985 the International Commission on Math-

_I_-- -_^,. p . ..- -..

H. Burkhardt and R. Fraser: Overview 1

An Update of the 1985 Strasbourg conference: 11

L. A. Steen: Living with a New Mathematical Species 33

S. B. Maurer: What are Algorithms? What Is Algorithmics? 39

J. Stern: On the Mathematical Basis of Computer Science 51

K.-D. Graf, R. Fraser, L. Klingen, J. Stewart and B. Winkelmann:

S. B. Seidman and M. D. Rice: A Fundamental Course in Higher Mathematics

B. Cornu: Teacher Education and Training 87

B. R. Hodgson and E. R. Muller: The Impact of Symbolic Mathematical

M. Mascarello and B. Winkelmann: Calculus Teaching and the Computer.

D. Tall and B. West: Graphic Insight into Mathematical Concepts 117

Annotated References 124

---,-._ 1.L ..*

are now numerical and, particularly, graphical, al-

._ -- v-e- .+Aeaw ..--.-- ._ .-- -. - .__--_ _~-_-... ...-~-

but much research evidence (see Bell et al, 1983) (b)

shows that the skills acquired are not reliably re-

mathematics you know could help you to tackle the velopment .

widely recognised as important in education, but mum)

-* ___- .---- - .--. -_^- .

THE EFFECT OF COMPUTERS ON MATHEMATICS

demic Press. Wolfram, S. [1988]: Mathematics, Reading, MA:

THE IMPACT OF COMPUTERS AND COMPUTER SCIENCE ON THE

2.0 The Changing Science of Mathematics

Lynn Arthur Steen

REFERENCES Jaffe, A. [1984]: 0 r d ering the Universe: The Role of

Ralston, A., and Young, G.S. (Eds.) [1983]: The

Input rn, n [integers >= 0]

For instance, Figure 1

By induction, it should be clear that fi may be

Input n, unit line segment AB

Example 7: Numerical Solution of Equations Example 8: Sequences of Heads and Tails

Input (none) buttons, approach (4) is available while approach

Manber, U. [1989]: Introduction to Algorithms, A

Maurer, S. [1984]: Two meanings of algorith-

Maurer, S. [1985]: The algorithmic way of life is

Maurer, S. [1991]: Proofs and algorithms: a reply to

Maurer, S. and Ralston, A. [1991]: Discrete Algo-

Mines, B., Richman, F. and Ruitenberg, W. [1988]:

NCTM [1989]: C urriculum and Evaluation Stan-

..- -._ ..I II--

means that, if execution terminates, the final for- Conclusion

Robinson, J.A. [1965]: A machine oriented logic

__ ._._ ..--..-_1__, -*----

tween members of the community. ploration, this software is independent of country,

knowledge and would be impossible to answer if you Autocalc Options

44 1. Helen decided to add one to the tens and then

_ _ . . ..___. -. . . - . .._-_-_ - -_-__ - --- .-UI

that begin t,o expand their understanding of num-

Geometry SYMMETRIC TURTLES (Graf, 1988)

shown that the kaleidoscope can be a window to

A new situation occurs if we start reflecting a tri-

. the viewing of a very broad spectrum of complex

Function plotting software in acceptable qual-

las for the logistic function, the growth behaviour

School Mathematics Curriculum 67

25 pupils in the class. Therefore, often theoretical

a nnite numoer of unltormlv dl.d,rlhl:tPd fnppllrlh

__ --..- .-._-. .--.