Discussion
Hmmmmm... Having skimmed all the dependencies I'm planning to use in a dumbphone (if I go through with it) & having recently discussed how I'd reimplement my own USB stack...
It strikes me that the one dependency I'm looking at (Tiny USB & its forks) which is obviously tainted... is also arguably the silliest!
Why haven't we devs automated conversion of the point-of-truth PDFs into starter-APIs? I'm not finding anyone who has at least extracted the tables into a more machine-friendly form.
When it comes to my laptop I don't feel its worth me personally stripping out any project which touches LLM-toolingfrom my laptop. I'd be driven mad! Besides I trust @elementary to do as a good a job, if still imperfect, as anyone.
But when I'm wanting to get away from modern dev-stacks & build a useful bare-metal system... I'm tempted to not put any such projects in there to begin with!
Even if the relevant hardware vendors like advertising "AI".
2/1!
I recall that the other week I find this project in my feedreader: https://yingtongli.me/git/pdfi
With this it looks like it might be reasonable for me to aid others innovating their own novel computing stacks in building their own USB APIs! With a little deterministic (Python) automation.
This is probably overkill for my own efforts, but given how disillusioned I am... I want to help in this way!
3/1 Truly fin! Until I've built it.
Edit: I poked around, I'd need a different tool.
Install bonfire.cafe
Get the full app experience
