This is a little experiment with Markov chain text generation. The code is from shabda (with a tweak to always end with a period), and the corpus is most of the text I posted to the WELL during the time I was most active, mostly from 1997 to 2000.
I find that I'm just not entirely sure how they work, but I have a PB 2400c, and a PowerBook 2400/240, and I am a very good card.