I would like to let you all know that I am yet another who is anxiously awaiting the release of this client for the linux platform. Good job on the work done so far.
While I have no clue about the design struction of TeamSpeak 2.0, I would like to offer some suggestions, in hopes that they might be incorporated in later versions[commence the flaming now].
I would offer to suggest that the GUI be seperated from the client, I know you have to consider that Windows is the primary target platform, but stay with me on this. This would provide a number of benefits.
The first benefit would be, the ease of port. No need to worry about an extensive GUI development cycle, just the command line version would be all that's necessary. As a matter of a fact, it occurs to me that since the Linux server version is already command line, maybe it wouldn't be too much of a stretch to incorporate the client capabilities into the server. Let me clarify, If I wanted to connect to a remote server, I would initiate a server of my own which would be told to relay only. Ideally, relaying from the client portion.
Second, you could ease some of the "MAKE IT OPEN SOURCE" intensity. Why/how you ask? Well, all the OSS peeps really want is the ability to customize the apps, for the most part. While TS2.0 is looking sweet, I think we'd all agree that it'd be nice to replace the GUI, of any app, with a flavor of our choosing. Well, if you release a command line only version of the software, people would be forced to write their own GUI's for it. That is not an exception in the Linux/Unix world, but generally the rule. You could focus on supporting/debugging the command line version only and not have to contend with who's running what window manager/desktop environment. People would write numerous GUI's for it. I guess the only thing necessary would be to release a TeamSpeak 2.0 library, some header files and an API. That wouldn't make TS2.0 open source and you'd shut a lot of people up

.
The last option that I would suggest is plugin support. That's a gigantic general suggestion and I'm unqualified to specify what exact plugin support should be defined as, but I'd imagine the possibility for people to utilize OGG compression/streaming, and other codecs.
Using these techniques, TS2.0 could still, remain out of the open source arena, enable people to customize it as needed, reduce porting and support issues, and remain legally licensed as it is now. It would be a license violation to write a wrapper/plugins and try to resell them, as it would be to hack an EXE and change some information just to claim it as your own. I don't think the Linux peeps are out to steal/profit/plunder, and I don't suspect that you think that either, I just think it's the ability to have a choice that they're after and the suggestion I've laid out would bring that possibility to reality.
Again, I reiterate, I don't even begin to assume that I have enough knowledge about TeamSpeak to suggest that these things are possible, they are just suggestions.