Back to corpus
research noteexperiment writeup candidatescore 24
656,000 Lines of Code for a Script Most AI Has Never Seen
Here's a number that hit me when I ran `wc -l` across the project: 656,000 lines. Python, Swift, TypeScript. All of it for a single writing system that most language models handle worse than random noise.
Full HTML reader
Read the full artifact
Extracted abstract or opening context
Here's a number that hit me when I ran `wc -l` across the project: 656,000 lines. Python, Swift, TypeScript. All of it for a single writing system that most language models handle worse than random noise.
The script is N'Ko (ߒߞߏ). You've probably never heard of it. That's the whole problem.
In 1949, a self-taught linguist named Solomana Kante designed N'Ko from scratch. He spoke Manding, a language family with 40 million speakers across West Africa. Bambara in Mali, Dioula in Cote d'Ivoire, Mandinka in The Gambia, Maninka in Guinea. Arabic script and Latin script both existed for these languages. Neither fit.
Arabic doesn't capture Manding's vowel distinctions. Latin doesn't encode its tonal system. The word "ba" means mother, goat, or river depending on tone. In Latin script, all three look identical. In N'Ko, each has a different mark.
Kante built something precise. 27 base characters, each mapping to exactly one sound. No silent letters. No irregular spellings. Explicit diacritical marks for all three tonal levels. It reads right-to-left. The name means "I say" in every Manding language.
Promotion decision
What has to happen next
Attach run IDs, datasets, metrics, and reproduction commands.
Why this is not always a full paper yet
Corpus pages are public-safe readers for discovered workspace artifacts. They are not automatically final papers. A corpus item becomes a polished paper only after the editable source, evidence checkpoints, references, figures, render path, and release status are attached through the paper schema.