This project documents my implementation of Jumbo, a program originally written "mostly in 1982" by Douglas Hofstadter, who used Lisp, an elegant weapon for a more civilised age. My real project, of course, is not to write a really abstruse way of finding anagrams that aren't even required to be English words, but to kick off the development of (my version of) the architecture as a whole. I'm starting with Jumbo because it has the fewest moving parts and out of a vague sense that ontogeny should recapitulate phylogeny. Because I'm Han Solo to Hofstadter's Obi Wan, I'm writing it in Perl, the common blaster of programming languages.
Code at Codeberg: https://codeberg.org/Vivtek/AI-FARG-Jumbo
You'll also need my terraced scan implementation AI::TerracedScan and the framelet library.
Project status: Halfway done at "Jum" stage while I build some more functionality into the terraced scan engine.
Intermediate report: TR2: Jumbo Revisited, Part One: Jum
More detail:
- Explainer about the Jumbo domain
- Background on how the terraced scan works at the Jumbo stage (there will be more later)
- A bit about developing the Chunkabet, the syllable-quality database that drives the Jumbo process
