But I was thinking more of something like the Nostr Protocol repo. That's code that doesn't even need to compile. It's just a spec.
Write the protocol for notes in notes.
Discussion
That feels like, more, using the wrong tool for the wrong job though. Repos do have sweet spots when it comes to huge, deeply nested structures of code where you probably will have changes that span multiple files in multiple places yet just make sense when applied together. Keeping something such as specification documents in a repo ... seems okay'ish but not the best choice for sure.
For this, I agree. For just documentation, repos are overkill. But I thought the initial post was about OS repos, not protocol specs. And OS repos do contain some 30 mln SLOC parts, something one can't easily post onto wikis, let alone here.
Not operating system, open-source.
nostr:npub163gcvh4dwwqm4yp2y7355tu9s7e6pzmqlcl3p78m7vm52fq7ej9s0g40f6 I guess I should have written FOSS.
Ah! Probably, yet FOSS repos still sometimes contain millions of SLOC. ;)
Yeah, but most are tiny and would get wider usage, if they lowered the barrier to change.
Lowering the barrier to change still would solve little, because most people only want to consume. Besides, contribution to specs is one thing but software code is a different thing. Some quality control still needs to be in place, to avoid malicious or badly performing code. Because, for instance, anyone can learn to write JS, but few can learn to write *good* JS that doesn't make a quad-core ARM feel like 6502. My point is, the problem is not with repos, the problem is with people. On both sides.