@moira Relays. Nobody is currently running one because nobody needs to, but people have spun them up locally and done proofs-of-concept with them to show that it can be done.
Top-level
@moira Relays. Nobody is currently running one because nobody needs to, but people have spun them up locally and done proofs-of-concept with them to show that it can be done. 10 comments
@MikeBeas I suppose I want to say: _I'm not rooting against them._ I just think that they're never going to be the kind of distributed that AP can be and that creates a particular set of types of problems that can't be handwaved away, once of which being the _reliance_ on Relay. Thanks for the update on portable identity. Assuming it's still blockchain-record based (as it was supposedly going to be) maybe it'll be the first actually _good_ use of that technology. (2/2) @moira It may be a lot of money for you or me, but it’s relatively inexpensive for someone trusted like Cloudflare or the Internet Archive to run. The intent was never for everyone to run their own separate relay, or for individual users to manage them. @moira The thing is, if bsky the company is behaving well, there's really no *need* for a second relay right now. if they go rogue, having the ability for the community, or some other org, or anybody else to stand up a replacement is a huge win. even just a partial relay for specific types of events, which the devs have said is fully supported and something they would expect instead of full relays, would be useful. but there's no need to use the fire escape unless there's a fire. @MikeBeas ...you say that like it's _easy_. hooooooooly shit it's not that will be an _adventure_. It will be much harder and take much longer than I think you think. But at least it will be possible, so at least there is that. @moira luckily it doesn't need to happen right now, but i suspect if it did, you would probably see a coordinate effort in the community to get one up and running as quickly as possible. there are already a number of apps besides bluesky using the protocol, so people do have an interest in keeping the network alive. @MikeBeas I'm sorry that "identity still centralised" was kinda mean on my part. I mean, I'm not wrong, but in real terms it's basically "DNS for AT Proto Social Media" and I just look forward to all the old jokes about DNS being repurposed for use whenever something with PLC is fucked up. xD @moira dns is a very good system! lol but technically they do support did:web alongside did:plc. you have to set it up yourself, and because your did is your permanent identity, you can't migrate between methods, but it is supported and there are a few people in the production bluesky network using it right now. |
@MikeBeas Oh, okay, I thought they had gone to some other name.
I know why they set up the relay based system, and I understand their reasons for it, even if it feels kind of like Weirder Bigger Token Ring to me. I appreciate the arguments for things like preserving sequentiality and inter-instance reply consistency (something we have now here too).
But I still think they're a mistake _because_ of the structural requirements in running one at scale, which _will_ discourage others from being set up even if it is in theory possible.
They require large money. Inevitably. And that just leads back to centralisation again. Yes, I know, in certain ways it's more efficient, but to me it feels like just moving the football again, Lucy.
(1/2)
@MikeBeas Oh, okay, I thought they had gone to some other name.
I know why they set up the relay based system, and I understand their reasons for it, even if it feels kind of like Weirder Bigger Token Ring to me. I appreciate the arguments for things like preserving sequentiality and inter-instance reply consistency (something we have now here too).