Beruflich Dokumente
Kultur Dokumente
5.
Relevant Systems & Root Definitions
A relevant system is one which is thought to be
helpful in learning about the situation - for any
situation there may (will) be several possible
relevant systems
A Root Definition is the name of a relevant
system.
The core of a relevant system is the
transformation it performs
Input T Output
Transformations
- The input must be transformed by the
process and the output must be a
product of the transformation
e.g. for a public library
6. Stock 5. Classify
shelves books
The Need for Monitoring
How do you know if the system is working?
6. Stock 5. Classify
shelves books
6. Stock 5. Classify
shelves books
8. Define
Performance 10. Take
Measures Controlling
Efficacy Action
Efficiency 9. Monitor
1-6
Criteria? New
'what's'
Alternative
'hows'
Implementation
The implementation process itself can be thought of as a system -
using SSM, where expectations are the only real-world analogue
SSM does not seek to address specific implementation techniques
Implementation
Main types of change
1. Structural
– e.g. organizational groupings
2. Procedural
– i.e. ways of doing things
3. Policy
– goals and strategies
4. Attitudinal/Cultural
1-3 relatively easy, 4 rather harder.
Danger in setting out to change attitudes
– should be 'by-product' of the overall process
Still ultimately based on a rationalist model
of human behavior
Rules for SSM
Tactical Rules
1. Each stage , 2 - 6, has a defined output.
Stage 2 - Rich pictures, Relevant Systems
Stage 3 - Root Definitions (CATWOE)
Stage 4 - Conceptual Models built from Root Definitions
Stage 5 - Agenda for possible changes derived from
comparisons
Stage 6 - Agreement on desirable and feasible change
2. Conceptual Models should be derived from Root Definitions and from
nothing else
3. Conceptual Models should be checked against Root Definitions
4. Conceptual Models are not descriptions of systems to be engineered
5. Don't look for systems in the problem situation - the systems are created
as (conceptual) tools for learning
Rules for SSM (2)
Strategic (a few pointers)
1. Resist temptation to impose structure onto
problem situation
2. Annotate Rich Pictures tersely, if necessary
3. Consider social roles & norms in the situation
4. You can have two versions of Rich Pictures -
public & private
5. Relevant Systems can be Primary Task or
Issue Based - consider both (look for them in
the Rich Picture)
6. Develop several Relevant System in parallel
7. Iterate! SSM is not prescriptively linear -
loop around stages 3 to 5.
SSM & Information Systems
- Information Systems are Social Systems
– for some people.....
- SSM is designed to deal with the complexity of social
situations
- SSM should therefore be suited to thinking about
information systems
- Which of the huge possible number of information systems that be
could develop, should we?
Information = data + meaning in a context
- Why do we need I.S.s?
- to serve real-world action
- The IS must therefore include the action of attributing
meaning to data
SSM & Information Systems (2)
An Information System has a least two parts
- Data manipulation is a machine activity
- Data transformation is a human activity
Transformation
Manipulation
SSM & Information Systems (3)
- Attention must be paid to the purposeful action the IS is supporting
- and the meanings that make those actions meaningful
& relevant
- Need to understand how actors conceptualise the situation (Ws)
- Enables construction of human activity systems for the situation
- Of each activity ask
- what information is needed to support this activity?
- sources, format, frequency?
- what information is generated by this activity?
- to whom, in what form, how often?
- Can then consider data structures to embody information
- Can then design data manipulation processes
SSM & Information Systems (4)
SSM can enrich and enable Systems analysis
and design
- particularly information requirements
determination
Gets away from assumption that problems are
trivial
Future Developments
An attempt has been made to incorporate SSM
into SSADM
- wrong way round - hard systems are a
special case of soft systems
Also attempts to use SA tools in conceptual
models, and also to use other models as
'templates'