I think I'm going to try my hand at the topcoder competitions.

I've been spending the last few days catching up on homework and working on WinTarBall. I was planning on advocating .tar.gz and .tar.bz2 once I released WTB 1.2, but I've discovered a few nasty things about .tar (and .zip for that matter). First of all, POSIX tar only allows filenames of 100 characters or less. To facilitate longer filenames, GNU tar uses a nonstandard extension, which breaks POSIX tar. Secondly, tar stores filenames in ASCII, so you cannot archive files whose names contain characters not in the ASCII character set. Zip files also suffer from this problem.

An obvious solution to these deficiencies is to create another archive file format, but do we really need yet another archive format? I'm still undecided about the best course of action, but as the computing industry moves towards internationalization, we will need a way to archive files with non-ASCII names.