graph based incremental interpreter

[jimy-byerley]

The current interpreter often fail to properly save long scripts, causing a lot of complete reexecution and visual artifacts
It also doesn't help with modeling divided accross several functions because most of the time functions are defined quit far before being called, cause the code starting from its definition to be completely reexecuted even if not related

It is also a pain to switch all the time between functions levels (and each time reexecuting the whole function content) just to visually check local variables

A new interpeter is needed:

• graph based, to ignore code not depending on non-altered code even after the earliest modified lines
• multiscope, exposing functions local variables even when running the global scope code
• updating the view during computation (more interactive)
• caching functions local variables

a big work in perspective

[jimy-byerley]

should also address #17 #14 #16