> Without checking, that was in the spec for converting the chat to muc > case. I don't see the harm in leaving it there, but I think it'll come > down to another argument about the impossibility of collisions on a > guid.
That was indeed the main discussion last time. Pros: it keeps the client very simple (it just has to query a name), and leaves out all the logic for GUID generation and collision detection. Others were arguing that a SHA is simple to generate, and would never collide in pratice anyway. I'm slightly leaning towards the first proposal, as I'm not sure I could sleep at night knowing that in theory, my client can give a weird error. But maybe i should stop being a purist, and accept that the clients have much bigger issues than that ;-) cheers, Remko