One more thing on the Debian OpenSSL bug. It inspired Web developer Bo Allen to look into the randomness of the PHP rand() function. He compared it with the results from random.org, which uses atmospheric noise as a random seed. The result is a visually clear example of randomness and not-so-randomness. Read the blog, you'll see what I mean.

Allen's test makes me think that someone (no, I don't have the time) should do the same with other programmatic random number generators.