Ok, today was awesome.
Indigo is... exactly what I need^H^H^H^H^H^H anyone who's programming service oriented system needs. I can't even put it into words how happy I was as the serialization session progressed. By the end of the session a great weight was lifted off my shoulders. Oh man, when Doug explained that the serialization architecture is now one model and that that model is the XML Information set... well... let's just say that I'm still smiling.
WinFS is everything I expected. It model is very similar to O/R mapping layers today. I've designed two versions of an O/R mapping layer and the latest version, which is .NET, bares a striking resemblence to WinFS. The reason I'm so happy about that is because now I can go back to my development team next Monday and spend about ten minutes explaining the analogies between the two and they'll basically know how to program WinFS. Awesome.
Now, back to reality. The “good” Indigo stuff really isn't coming until Longhorn. Although there are some improvements coming in the Whidbey bits. WinFS? Well there's no doubt we're waiting 'til Longhorn for that. So the negative perspective that some people are taking is: “I love to see this stuff, but it's useless because I can't use it for the next three years.“ To some extent, I feel the same way, but then I realize something: We, the developers, have a real opportunity here. We're seeing the APIs in their toddler stages. We still have a real opportunity to shape the outcome of this stuff. If we start playing with the bits and come up with constructive criticism (whining isn't gonna help) I know Microsoft will listen. It's either that or forget what we've seen this week and wait until 2006 and potentially be stuck with ugly, difficult, unintuitive designs.
Speak now or forever hold your peace.