Wave is Google's plan to provide interactive content over XMPP. It's a step in the right direction from HTTP, but XMPP still has its own weaknesses. Discussing some plans to implement a client capable of matching Google Wave's hype.
"So it's another chat client."
"Why would I use that over Pidgin?"
"If Pidgin will let you do data entry in online games and other interactive channels using a web form instead of a command line, none."
"Put HTML, CSS and JS in the content window and you have an interactive GUI to do Wave-like collaboration or AImapper-like order entry in games."
"I can see why this would need its own protocol."
"Google is trying to do it over XMPP. That's what wave is. It would work, too, but XMPP does presence via unicast. That makes traffic maintaining contact lists proportional with the number of users times the size of their buddy lists."
"Well, that and the only client supporting wave extensions to XMPP is Google's web-based wave client."
"That and XMPP doesn't do multi-user chat well."
"Or binary data."
"So, yes, you need a text based protocol that handles binary data asynchronously and provides programmable contexts for content and multicast routing. So, yes, a new protocol. :-)"
"So, PSYC will do the job, but the protocol itself is in development."
"And we want to be one of the beta testers."
"Oh, yes, we do."
"It's either that or follow long polling and AJAX and Google Wave along the same lines that table-based HTML formatting went."
"The de-facto standard, because nothing else really works?"
"Right. All talk about standards is useless unless the proposal actually does the job. This is the only thing that will do the job that's on the boards now."