Not sure I like this approach, I would build a proper client
This 1.23 BTC bounty for a nostr Github replacement looks really yummy 😋:
https://bountsr.org/code/2023/01/19/nostr-based-github.html
Here's my plan:
1 - Propose a NIP introducing 3 new event kinds (repo, issue, PR). Each event has an eventID and has a "parent" field pointing to another event. The root event in the tree is the "repo" event kind.
2 - Fork gitea/gitlab
3 - Add a nostr client library that can talk to
4 - Remove all server side code that queries a centralized database
5 - Replace it with event queries
6 - Cache it so it is not super slow
7 - Replace user database with a nostr login
8 - Self host a demo client
9 - Claim bounty to #[0]
10 - profit?
Discussion
There’s 230k waiting for you to build it…
Why reinvent the wheel when it comes to UI?
Wheel so old. Can't we have a sphere?
This is always proposed (fork an app and add nostr to it) and is almost always a bad idea. It would deserve its own clients and web uis without a centralized DB. The goal is to not just rebuild another github, but instead to do it right on top of nostr.
That's exactly points 4 and 5 from my OP.
I just want tl fork the frontend part not the backend.
I see no value in starting from scratch when it comes to that. Sure a bunch of useless crap has to be stripped down.
No the plan is the keep git. So it’s not clear to me what nostr adds?
The other stuff. Code review, issues, pull requests, patches, etc
Why so much resistance to just doing it all in nostr?
Seems 100x easier than combining git + nostr.
I think it worth a try to fork gitea. Every git service has pretty much the same UI and I think that convention should be followed.
I don't like it either, there is no need to re-create Git, it already exists. The problem is how to decentralize a website that saves user data. ZeroNet is worth a look, it does have a github clone working inside it