My original plan for 0.4.5

[utensil]

This issue is a placeholder for my original plan for 0.4.5, which will be postponed to later versions, but useful as a reference for forming a roadmap.

Moved from #17 (comment):

I think I'll be working on the following features in 0.4.5:

1. Performance: GAlgebra is extremely slow for certain coordinates and metrics, primarily caused by eagerness in the calculus stuff and simplification, so the main idea is laziness, so higher dimensions can be supported
2. Better support for FGA and CGA: it's already partially supported but I need to extensively verify them and some complications are expected
3. SymPy 1.4 support: I need to reimplement collect in SymPy for non-commutative symbols, that's quite some work, see Some tests are broken for SymPy 1.4 #31
4. Doc improvements: I want to write a true tutorial, weaving the examples together; also the docstring in Python code are mostly ill formated;
5. Examples fixing: examples in doc/python are not verified; some of them have serious correctness issues yet to be fixed.