Yesterday, I wrote a post about generating completely random names/words from a character set. Well, out of boredom, I’ve refined my algorithm, and, somewhat to my surprise, it now produces names/words which are not only interestingly alien, but actually phonetically consistent:
What I did was to modify the original algorithm so that, rather than just sticking together random letters cushioned by a vowel every other letter, it creates a list of possible two-letter phonemes (in addition to an occasional sprinkling of extra characters), and builds the words from those phonemes. This is good in that it produces pseudo-consistency in the words I generate (it’s not real consistency like you find in language, since there are no rules (not yet, anyway) for how phonemes can and cannot fit together). The only downsides are that: 1) the generator now very rarely produces three-letter words, and 2) the generator never produces inanely amusing things like “FUKER” and “CUJO'” from the previous list.
If I ever optimize it to a degree with which I’m satisfied, I’ll be sure to post the Python source code.
P.S.: Sorry about the lack of real content lately. The holiday season’s made me sluggishly overfed and rather lazy.