How to...

do differential geometry
One of Geometry Expressions' unique features is the ability to use generic functions. When you use generic functions in a parametric curve, you have a starting point to do differential geometry. For example, the evolute of a given curve is the envelope of its normals. To derive the equation for the evolute we:

First create a parametric curve, and set its x coordinate to be f(T) and y coordinate to be g(T). Now put a point on the curve and constrain its parametric location to be t. Select the point and the curve and construct a tangent. Select the point and the tangent and construct a perpendicular (this is the normal). Select the normal and Construct/Locus to create the envelope. Now select the envelope curve and Calculate/Parametric Equation to display its equation.

By creating a general curve in this way, along with the tangent, envelope and locus constructs, you can explore many results in the differential geometry of 2D curves.

How to...create an involute curve

The involute of a curve is the locus traced out by the end of a piece of string as it wraps around the given curve. Here's how to create an involute in Geometry Expressions. (In the picture, our given curve is a logarithmic spiral) First create an arc on the given curve.

 Now constrain the parametric location of the arc's endpoints. One should be a constant, the other variable (we've used pi and t). Measure the arc length. Create a tangent to the curve at the variable end of the arc. Place a point on the tangent and constrain its distance to be a constant minus the arc length. The constant represents the length of the string. The locus of the point on the tangent is the involute curve. For the logarithmic spiral, the involute is simply a 90 degree rotation of the original curve.

How to...measure the area between curves

There is a trick to measuring the area between two curves. You can create an arc lying between two points on any curve. But (with the exception of circles) GX will not let you put a point on the intersection of two curves. So we resort to the trick of joining two arcs by line segments then making the ends of the segments lie on the intersection points. Here's how: First create an arc between two points on the first curve and between two points on the second curve. Join the ends of the two arcs by line segments and create a polygon from the two arcs and the two segments. Now use Constrain / Proportional to specify the parametric location of the vertices on the two curves so they lie at the intersections.

Where to take a rugby place kick

The sport of rugby is a lot like American Football, but when you score a touchdown (they call it a try) you take the extra point kick (they call it a conversion) from a place in the field in line with the place the try was scored. So if the try is scored close to the sideline, the kick is taken from somewhere close to the sideline. If it is scored in the middle of the field, the kick can be taken from the middle of the field (Figure. 1).

Figure 1: Rugby extra point kick is taken from a position on the field in line with the location of the touchdown If the goalposts are width g, and the try is scored distance d outside the left post, what is the best place to take the kick? This will turn out to be a problem within the following topic areas: Calculus optimization Conic curves Finding the best place is usually mapped into maximizing or minimizing something. But what should we be maximizing or minimizing? What about distance, where would the closest point to the goal posts on the permitted line be? Would this be a good place to kick from? If not, why not? So if not distance, what else could be optimized? How about angle? Presumably it would be best from a point of view of accuracy if the angle between the sightline to the near post and the sightline to the far post were as large as possible. Use Geometry Expressions to get a formula for the size of this angle in terms of the width of the goals g, the distance between the near post and the location where the try was scored d, and the distance x from the goal line (Figure 2).

Figure 2: Angle made by the goalposts at the point of the kick Lets assume that d is positive (that is, the try was not scored underneath the goal posts.) What can you say about angle BAD? What can you say about angle ADB? Have a look at a graph of the arctan() function. What is its domain? Is it monotonic over that domain? Is it true that the maximum value of the angle comes at the maximum value of the argument of the arctangent? Use calculus to find the value of x which maximizes this quantity. When you have found your solution, you can plug it back into the diagram, replacing x by the solved value:

Figure 3: Solution configuration Try dragging the point D (You may need to lock g, to keep the width of the goalposts fixed!). We have created an algebraic answer to the problem, and we have a geometric diagram of a single solution. How can we create a diagram which shows the solutions for all values of d? Create the locus of the solution point as d varies. We will want to look at the equation of the locus in due course, we should fix our drawing in the coordinate plane first. One way to do this would be to remove the length constraint on the goal and instead specify the coordinates of each goal post, (Figure 4).

Figure 4: Locus of optimal kick points is a hyperbola What kind of curve is this?

Advice for a rugby kicker

We could tell a rugby kicker to place his ball on a specific curve, however, it is unlikely that in practice he could follow that advice. Is there some way we could approximate the curve and give the rugby kicker a rule of thumb to work with? Can you find an asymptote to the curve and plot it in Geometry Expressions?

Figure 5: The hyperbola asymptote is the line Y=X Is the asymptote a good approximation of the curve? What practical advice would you give to a rugby kicker?

Further Questions
What about the situation where the try is scored between the posts? What is the maximum value of the angle, and where is this attained? Would this be a good place to kick from? What factors do you think would influence the placing of the kick in this situation?

Notes
The derivative of the argument of the arctan function is:

Solving for x gives: