For a class project (due at the beginning of December), I have to work with somebody to make something that demonstrates a few of the following: "threading, networking, memory management, file management, inter-process communication, synchronization." It's for a class called "Systems Programming."

Does anybody have any ideas for something interesting we could do, even just very general ideas? I'd prefer that it's something useful to people and hasn't already been done, and that doing low-level "threading, networking, memory management, ...." on our own would make sense beyond educational purposes (so even if there are higher-level tools that generally make these tasks easier, there would be good engineering reasons for not using them for the project).

Write a parallelized raytracer or something cool like that (this could involve IPC + synchronization + networking).

Write something cool for Doors CS + CALCnet! I'll have to think about this more to come up with some ideas, although I like the direction Elfprince is going in.

A virtual machine includes some or all of those, depending on how far you take it.

What language are we talking here? If you want to hit as many of those as possible, perhaps some sort of collaborative editing? A a version of google docs sort of thing (and perhaps not browser-based but a desktop app)

IMHO the raytracer thing has been done to death, not very creative. It's a super well understood problem with fairly straightforward solutions (raytracing isn't particularly difficult).

Make one capable of rendering fractal/procedural surfaces - a mandelbulb raytracer (or something like it) would be cool.

I think it can be in any language (and if my partner and I don't know it already, we shouldn't have a problem picking it up). But, the language we're going to use in class is C, so it might be slightly easier to use C.

Thanks for the suggestions so far (keep them coming). Some sort of collaborative editing seems interesting, although I'd want to figure out how to make it actually useful to people and have the desktop (instead of browser) implementation actually make sense from an engineering standpoint. I'll also look into the other suggestions and see where I could do something new and useful.