i wouldn't really rely on it for an important project. it's too early.

the wrapper already works nicely but there still are some areas that aren't very tested (like the material system) and some stuff i still have to think about (like game_save()/game_load() - i think saving the script state could probably be done with pickling somehow).

i don't know how long it will take me. it would be helpful if more people played around with it and gave feedback about any problems but unfortunately there aren't many python users here.

...
i also prefer python over lite-c in most cases. lite-c isn't that bad but often i find that c forces you to work on a too low level and the debugging features are a bit lacking. debugging can be very annoying with lite-c. there isn't even a stack trace.

...
by the way, the wrapper is written in cython. it kind of is a python to c compiler. so you wouldn't have to directly work with c source code but you still need good c knowledge for writing a wrapper. you need to know how pointers and so on work.