Yeah, that’s a problem with a lot of FOSS passion projects. We devs kinda like writting code, but not really documenting it. Hopefully with the influx of devs helping that will improve
It’s a pretty major pain point for a lot of local libraries, too. I’m so glad I have the option to dig through the source code of functions if I can’t figure them out.
I don’t really like all the LLM hype, but I’m hoping that documentation will eventually be generated by some open source model, with human verification
Finnegans Wake makes more sense than Lemmy API docs. Even calling it “documentation” is a stretch.
I literally had to clone the Lemmy git repo and read the source code to find the implementation of an API endpoint and see how it worked for a script that I was writing.
These days the standard is to create an API Doc out of a OpenAPI document generated from the code itself.
Someone will probably contribute to it at some point.
I just wished the Lemmy API docs were better lol.
Yeah, that’s a problem with a lot of FOSS passion projects. We devs kinda like writting code, but not really documenting it. Hopefully with the influx of devs helping that will improve
More devs that don’t want to write code?
Edit: damn, meant to say write documentation. We only want to code!
Well, more devs that have to sit down and figure out how the code works and then document it for future new devs
I don’t think that is contained to FOSS passion projects ;)
It’s a pretty major pain point for a lot of local libraries, too. I’m so glad I have the option to dig through the source code of functions if I can’t figure them out.
I don’t really like all the LLM hype, but I’m hoping that documentation will eventually be generated by some open source model, with human verification
Maybe we’ll eventually get the corresponding influx of tech writers.
I think that this line of reasoning becomes less and less tenable when things like Swagger exist.
It’s like almost every piece of software, period.
Finnegans Wake makes more sense than Lemmy API docs. Even calling it “documentation” is a stretch.
I literally had to clone the Lemmy git repo and read the source code to find the implementation of an API endpoint and see how it worked for a script that I was writing.
These days the standard is to create an API Doc out of a OpenAPI document generated from the code itself. Someone will probably contribute to it at some point.
Fediverse software API documentation is bad across all fediverse software.