So when Bluesky introduces a new feature or a breaking change in the protocol anyone downstream will find out when it gets pushed, maybe a little ahead of time when it comes in as a pull request. Bluesky goes live with the change immediately, maybe in a public beta channel, maybe straight to prod, depending on their testing setup. Anyone running a bluesky compatible server becomes immediately incompatible until they rush to implement the new changes. The best user experience will be had on first party servers, driving the vast majority of users there.
For a standard defined protocol, like ActivityPub for example, to introduce a change like that it would first go to the standards comittee where it would be discussed publically with stakeholders. The changes would be published and then all parties would begin implementations at a pace that makes sense to them. It’s like when you hear about new Wi-Fi versions several years before any devices actually support them. One group doesn’t just get to come out with some crazy new change that everyone else has to reverse engineer and then race to keep up.
What Bluesky is doing might be fine and make sense for their model, whatever that may be. I just want to point out that there is a difference and it drastically changes what the future of the service will look like.
Nope, I’ve seen nothing firm about them doing or not doing it, but OP asked about current information and at the moment they are open source and do not have a standards document.
As it stands I wouldn’t jump on AT for a project of mine and wouldn’t recommend it as a superior protocol on those grounds.
I wouldn’t recommend it either, because it doesn’t exist yet. But its extremely disingenuous to make claims about how it will work when/if it begins federating in the future and declare it EEE.
So when Bluesky introduces a new feature or a breaking change in the protocol anyone downstream will find out when it gets pushed, maybe a little ahead of time when it comes in as a pull request. Bluesky goes live with the change immediately, maybe in a public beta channel, maybe straight to prod, depending on their testing setup. Anyone running a bluesky compatible server becomes immediately incompatible until they rush to implement the new changes. The best user experience will be had on first party servers, driving the vast majority of users there.
For a standard defined protocol, like ActivityPub for example, to introduce a change like that it would first go to the standards comittee where it would be discussed publically with stakeholders. The changes would be published and then all parties would begin implementations at a pace that makes sense to them. It’s like when you hear about new Wi-Fi versions several years before any devices actually support them. One group doesn’t just get to come out with some crazy new change that everyone else has to reverse engineer and then race to keep up.
What Bluesky is doing might be fine and make sense for their model, whatever that may be. I just want to point out that there is a difference and it drastically changes what the future of the service will look like.
Do you have any sources for that? I’ve seen no indication they don’t intend to release the protocol as a standard and that is a pretty big assumption
Nope, I’ve seen nothing firm about them doing or not doing it, but OP asked about current information and at the moment they are open source and do not have a standards document.
As it stands I wouldn’t jump on AT for a project of mine and wouldn’t recommend it as a superior protocol on those grounds.
I wouldn’t recommend it either, because it doesn’t exist yet. But its extremely disingenuous to make claims about how it will work when/if it begins federating in the future and declare it EEE.