Some things are worth criticizing about PSYC.
- We took too long.
- Yes. We certainly did. We failed to attract much support.
- Why didn't we get more support?
- Because we are seriously lacking at marketing and because our website design sucks.
- Because we are taking too long. And because we want much more than them. We have two very nerdy bad habits: idealism and open source.
- XMPP is covering some of the things that we wanted more.
- Yes, many in fact. But we don't need to re-invent all those XEPs, we just route them over what we are good at. Our interserver protocol. Of course we'd have done this and that differently, but we can stick to the good things in Jabber.
- Where is the full PSYC specification?
- As stated at the end of TODO, as long as we rip the protocol into parts we cannot provide a full specification, still this site is full of best practice explanations how PSYC is working currently and a real specification is coming up.
- Why after so many years are we still ripping up the protocol?
- Because we are still experiencing a-ha effects. When you suddenly realize, the protocol should be able to handle arbitrary complex data (more like JSON, not XML ;)), should avoid escaping and it should be entity-oriented. And it should stop thinking so much in a request/response way and push decentralized state instead. Sorry for throwing techie words at you, but some of this is actually important. We managed to update the syntax to allow for these things, without actually making it harder to implement. That's cool.