Standardizing p2p software development collaborative platform:
ariard opened this issue · comments
If there is a list of Nostr-based and non-Nostr based experiences of replacing Github (e.g the radicle, we could start to agree on an endpoint format (probably a superset of bip341 npub and rfc5322 email address and few other data fields).
From a common code contributive endpoint format, we can certainly design new additional protocols, e.g user-assigned and authenticated default git repository hosted by Nostr relay or any other 24 / 7 peer-to-peer servers like.
Once you have user-assigned git repository, anyone can start to accept Nostr notes as review comments on its code proposal to change the hash-identified tip of a code repository. This review process can happen on everyone private or community Nostr relay.
Once you have local review process and git repository, you can apply some byzantine fault-tolerance consensus mechanism to convergently synchronize all the code proposal local review process. Any code proposal which at reach some review threshold aggregated from an authenticated code contributive endpoint can be make a candidate for synchronization. Synchronization can be clocked on period Bitcoin blockchain cycles or any other reasonably secure time protocol.
To sort among “signal-vs-noise” code change proposals a reputation algorithm based on a PGP web-of-trust can be leveraged. Anyone local git repository is free to apply more strict code policy validation. Trustworthiness of the process should emerge with time according to PGP security philosophy (see “web of trust”).
Anyone can apply its own collaboration policy on its “hosted” or “self-hosted” git repository (e.g see related : bitcoin/bitcoin#29570).
If further interest, I can propose a NIP or BIP and some proof-of-concept rust code.
I don't understand what you're proposing.
I don't understand what you're proposing.
I’ll propose nip or code more speaking.
If there is already experience of Nostr-based GH replacement, interested to look on them.
There is NIP-34 and a bunch of experimental clients. The experience we're aiming for is that of people running git servers on their own machines or on different providers, but able to interoperate easily by sending patches over Nostr.
See https://github.com/DanConwayDev/ngit-cli, https://github.com/fiatjaf/gitstr, https://patch34.pages.dev/, https://gitworkshop.dev/
The experience we're aiming for is that of people running git servers on their own machines or on different providers, but able to interoperate easily by sending patches over Nostr.
Okay interesting, same mental model. Thanks for the references.
I’ll close this issue and look more into NIP-34, before to see what can be improve.