Avoid confusing reuse of 'x' in spatialDistribution description

[henrikt-ma]

Cleaning up the notation used for spatialDistribution, plus stating that initialPoints and initialValues are parameter expressions (item 3 in #3640 (comment)).

Note that the change about parameter variability is not the one suggested in #3640 (comment), so this still remains to be done.

For the notation part, the most important improvement is to avoid using x as the name for the normalized position in the equations, since this was very confusing when the formal parameter x in the synopsis as well as in the pseudo-code is something else.

[HansOlsson]

The other changes look good, but I'm not sure if replacing $x$ by $\xi$ is that beneficial.

On one hand using x both in the differential equation and as input is a bit overloaded, but on the other hand we lose the connection between the x-coordinate as input and the $\xi$, and in particular state that $v=\xi'(t)$, and then describe $x$ as the integral of $v$. Effectively that combines to imply that $x=\xi$, right?

[henrikt-ma]

On one hand using x both in the differential equation and as input is a bit overloaded, but on the other hand we lose the connection between the x-coordinate as input and the ξ , and in particular state that v = ξ ′ ( t ) , and then describe x as the integral of v . Effectively that combines to imply that x = ξ , right?

Right. The relation $v(t) = \xi'(t)$ is crap, as ξ is an independent variable and not a function of time. This relation is now removed. That x is a function of time, while ξ is not, also highlights how different they are and why using an overloaded symbol gets very confusing.