Cleanup of Haskell model

[rlaemmel]

Looking at this:

The fold is not "data" perspective. It may count as "Run-time system" (new subfeature of "Dynamic semantics").

Since the event sequence is shown in the same piece of code, which is indeed "data", you would need two perspectives.

Anyway, the code is not well modularized in two ways:

• Why mix a general fold (which works for many inputs) with one specific event sequence?
• Can't you have a more polymorphic fold which also does not depend on the specific state type?