mod1_dynamic as composition

[fph]

Now that the support for composed functions in preimages is in good state, it would make sense to try and replace mod1_dynamic with the composition of the underlying map $T(x)$ and the "mod 1" map. This would reduce some code duplication, since what we are doing now in mod1_dynamic is essentially computing the preimages of integers.

In addition, I believe that thinking about this problem would lead us to write better interfaces to support composition and dynamics more generically (e.g., with different domains, or by a new dynamics type for the map like x -> x mod 1); it is a good model problem to have to think about what we would like our dynamics + composition interface to support.

I'm not going to work on this right now, but I'm just noting it down as a future idea before I forget. :-)