I would expect the Pygame-on-ctypes proposal to be a full two-months work by itself,
I read the page about the pygame-on-ctypes project (I like the Pygame idea too) but what the pygame crew think about it? I tried to ask on #pygame some days ago but they didn't seem really interested (it's preferable doing something *really* useful...)
but tackling smaller modules might indeed allow 5 or more of them to be done.
That's great.
About the modules choice, yes, it's mostly up to the student. There are many built-in CPython modules that we are missing, and thus having them would be nice, but on the other hand some non-built-in modules are just cooler and more widely used (I'm a Pygame fan :-).
I know (I remember something from your psyco-python-pygame presentation)
* _curses * _ssl * bz2 * fcntl * mmap * os (i.e. posix/nt, partially done) * readline (a topic by itself: write or find existing readline replacements in pure Python, also useful for CPython) * select * signal * termios * time (partially done) * zipimport * zlib
I'll take a look at them and try to decide which and how many. If you (or other PyPy dev) have some preferences (or priorities) let me know!
Feel free to propose any choice, from this list or not, here or on #pypy.
I'm often there. rhymes is my nick :) -- Lawrence http://www.oluyede.org/blog