There is some interesting discussion on the LaTeX stack overflow page about the challenge of detecting and preventing rivers during the typesetting process:<p><a href="https://tex.stackexchange.com/questions/4507/avoiding-rivers-in-successive-lines-of-type" rel="nofollow">https://tex.stackexchange.com/questions/4507/avoiding-rivers...</a><p><a href="https://tex.stackexchange.com/questions/29049/how-to-define-the-badness-of-a-river" rel="nofollow">https://tex.stackexchange.com/questions/29049/how-to-define-...</a><p>And if you really want to get into it, there is a rather detailed paper by Alex Holkner: <a href="https://gwern.net/doc/design/typography/tex/2006-holkner.pdf" rel="nofollow">https://gwern.net/doc/design/typography/tex/2006-holkner.pdf</a><p>Avoiding rivers becomes a rather non-trivial optimization problem. In Holkner's paper he found that it took ~1 minute just to typeset 1200 words. Some of his experiments took more than six hours to complete.
Since Christian Bök’s <i>Xenotext</i> was recently discussed here (<a href="https://news.ycombinator.com/item?id=43974005">https://news.ycombinator.com/item?id=43974005</a>), check out “The March of the Nucleotides”: <a href="https://x.com/christianbok/status/1064741910948995072" rel="nofollow">https://x.com/christianbok/status/1064741910948995072</a><p>In this case the edges of the river represent DNA base pairs: A on one side has always a T in front of it; the same with C and G.
Futility Closet had a cheerful, wholesome podcast for many years until abruptly ending it without much explanation. Glad to see they're still busy. Anyone know what happened? I always wondered.