Experiments with py5, I'm not sure how much further I'm going here (my python is v. rusty anyways). Idea is excellent, and I may need to to explore jupyter notebooks as a new programming paradigm. However things could be so much better with a radical re-write of processing.
Particular irritations:-
- Processing
map
is not polyglot friendly (should beremap
or similar). - PApplet is behemoth (py5 and ruby-processing projects have at least moved noise implementation out).
- PVector is a mess, overloading Vec2D and Vec3D functionality (I'm happy with my java extensions here).
- Processing4 is even more patched up than earlier versions, dead code and MacOS fixes abound.
- Inner classes are a mixed blessing (fortunately there's a meta-programming workaround for ruby-processing).
I'm not sure p5 static mode is good idea, apart from being terse, python mode should be pythonic. Which is why I will concentrate on class mode, or possibly module mode as seems suited to jupyter notebooks.