Right now, only the plan for M1 is fully set in stone, although M2 and M3 aren't far off. There will probably be a few more milestones after that too. The overall development I think will be split over two main phases. The first few milestones will deal with improving chat, while the second batch of milestones will focus on the GUI.
M1 is mainly under the hood stuff such as the rewrite of the widget code, implementation of widget containers which is very important for GUI construction, and implementation of context menus using the new widget system. Those things are more or less done, with menus needing a bit more polish. What's left to do now is the implementation of chat window filters that will allow you to dedicate a window to a certain conversation, such as shout or tells. The context menus is still blocking this for now until checkboxes and arrows for submenus are added. Also, it will mean you are no longer restricted to just two windows, one for system messages and one for chat. You can open in theory an unlimited number of windows.
M2 builds on that by implementing tabbed chat windows, which will be the focus of that milestone. This should be a much shorter milestone than M1, since the main rewriting of the code is done then and nothing major needs implementing besides the tabs themselves.
M3 then focuses on rewriting the chat windows. That makes it possible to implement proper fonts that can be resized, and will also allow text windows to be resized horizontally.
Really, the order of M2 and M3 could be switched if needed as neither are blocking the other.
After that comes the GUI phase, where the existing widgets get componentised to make full use of the container system. This is where the interface will have to change, so it might as well get improved at the same time. It should see the various elements of the interface get standardised, and much more code should get reused here too. Will probably involve quite a bit of code destruction too. Should be a pretty lengthy milestone there as a lot of things have to be done in one go, though maybe not quite as long as M1. After that comes advanced skin support where you can define the various elements of the interface in text files, which is an improvement over simply swapping bitmaps.
The GUI phase is probably thinking a little too far ahead atm. Originally the DWC was started on simply to improve chat as it was becoming unmanageable, but as the changes needed to do that was to rewrite bits of the client, it might as well go the whole hog after the chat improvement phase as then the work should already be done towards making the GUI phase a bit easier. Anyway, that should give an idea about where the DWC is heading over a fairly long period of time.
Thinking a bit closer to the present, the plan for M1 is to get the chat window filters implemented, then it will be more or less feature complete, then a beta can possibly be released if needed. There might need to be a bit of time to fix the rest of the bugs then and polish it off so it's fit to be merged to the trunk and released.
I can't give a clear time-frame on it yet. I don't think adding the chat window filters will take long, maybe 1-2 weeks. But before that the context menus need checkboxes and submenu arrows implemented which will probably take longer because coming up with a solution to that is a little tricky. Then there's RL issues plaguing me atm. Maybe a conservative estimate would be 2 months until it hits beta and sheds its vapourware coat.
Also, "Two Weeks" is getting old now, as old as "But can it play Crysis?". I'd rather leave it behind at Dai where it belongs.