Ah, but the really interesting parts aren't open source. (The server/client part.)

Perhaps someday.

I'm interested enough in Wave to consider reverse-engineering that portion. I think there are some once-in-a-lifetime opportunities here...

Yes, the server and client will be opensourced according to Google (which has a good history opensourcing their stuff).

And yep, some people can't wait so here is a server implementation (Python/Django):

http://code.google.com/p/pygowave-server/

They confirmed here: http://groups.google.com/group/google-wave-api/msg/b6649dcc5...

"Yes! The plan is to release a reference implementation of the client, as well as the server. "

I tried to do it myself and you pretty much have to write your own textarea implementation in Javascript. Wave as an idea is good. As a protocol I think is horrible for Javascript because either implementing the straight server-to-server spec is impossible on a regular textarea, or you need a beefy enough server to parse full-text states between keys for X many users.

While Wave is a good idea, and the hype is justified for a closed system, implementing it on your own system requires a deep, DEEP knowledge of Javascript.

What type of opportunities?