Git requires a human to resolve conflicts, a blockchain comes to a majority consensus. But yeah, I would use git and try to automate conflict resolution but it stands to reason blockchain can also solve the problem. If you have 10 people live-editing a text content git will be messy for example but you can come to a consensus on agreeing who made what changes and when between participants (assume p2p connection).
Git is too slow for the use case I mentioned you resolve conflicts when you merge to your repo but you don't establish a consensus allowing everyone to accept your conflict resolution without centralizing.
Establishing consensus is more difficult than a single person deciding what to merge, if you need a trustless system then you can add consensus on top of Git.
Yeah and speed up git to commit/push/merge very fast. Or use a blockchain. I feel like there is stigma against the term, shall we call it distributed leger instead?
You can call it a BitTorrent magnet for all I care, it doesn't change the fact that a distributed ledger only has so many uses. Git already has consensus mechanisms baked into it's architecture, and immutability is provided with GPG signing if you care enough about it.
Both of those features have been available in Git since the 90s. You're not revolutionizing anything by proposing a migration to an even less ubiquitous protocol than HTTP or SSH.
> You're not revolutionizing anything by proposing a migration to an even less ubiquitous protocol than HTTP or SSH.
I did not propose that. I didn't know git had a consensus mechanism built into it other than manual resolution of merge conflicts. A simple blockhain over https is easier for me than trying to get git to do something it wasn't designed to do (it works with files and repos, can you imagine having a commit+push+merge for every word edit in my example?) there are ready made libraries for this. The path of least resistance to acheive desires goals is what I would take. To each his own, blockhain can solve these problems outside of crypto.
Another big problem area is PKI as well. I'm sure you've seen stories about suspicious CAs, a distributed ledger of CA issuance and in the wild certificate observations is one legitimate goal that can be solved by blockhain (better than centralized CT logs that don't take ITW observations).
> I didn't know git had a consensus mechanism built into it other than manual resolution of merge conflicts
Git has cryptographic identity baked-in, so people can assert authority over their own contributions. That enabled mailing groups and IRC channels which encouraged discussion before arriving at a consensus and merging relevant files to the master branch. It's not decentralized, but it could just as easily be hosted as a torrent if that was something people cared about. However, decentralized development has historically been a farce, so most people use something like Github to control and moderate their repos. You're right, this is a "to each his own" situation, and there's literally no one on the "hosting Git via BitTorrent" side of the room. Moving to a less-efficient, more-convoluted process does not fix this. I've been following decentralized technology and the likes of Bitcoin for almost a decade, and the successes and failures of the space are self-evident.
> Another big problem area is PKI as well.
To who?
To the scary boogyman white-supremacist KiwiFarms NBC fearmonger types? Yeah, have fun getting anyone to rule in favor of them. To the rest of society, centralized CAs are a boon. Not only are they wicked-fast (good luck querying a blockchain faster than a DNS request), they go out of their way to prevent systemic abuse and domain-squatting in ways that crypto can't. You want to apply PKI to this space? Convince me that it wouldn't start another FTX-like situation, with powerful individual exerting economic power to centralize the market a-la crypto exchanges.
