.NET Reality Check...
I agree with Frans Bouma that there is a significant amount of hype around Whidbey, Yukon, Longhorn, etc, and it isn't all that useful for what we do today.
However, I think the work of the Prescriptive Architecture Group and the Patterns and Practices are good examples of non-trivial documentation and code that works with today's release bits. The PAG is working on today's problems and with today's bits.
The application blocks have been a tremendous help in many ways for applications and corporate frameworks we are developing for our clients.
http://www.microsoft.com/resources/practices/
I still find the discussion of the new stuff interesting, and I try to keep up so that when it becomes real bits, I have less to learn, and when I am building stuff today, I can think about how the nextGen stuff will be built, and design todays stuff in anticipation of that (to some degree). I find myself learning a bit more of the .NET Framework on each project I work on. And it has only been about six months since the last VB6 project I worked on.
SOA, for example, doesn't mean that I have to use Indigo. I can do SOA today, building framework pieces and components in a similar fashion to how Indigo is doing it. Those guys have already done a bunch of the design for me, so why wouldn't I leverage that? When Indigo arrives, I can adapt my pieces to use it instead of the “roll-your-own” that I built.
Mike
An interesting stat: there are about 28,000 method calls in the Win32 API, and about 184,000 method calls in the .NET Framework.