I don't think they are going to open-source legacy stuff that they want people to forget. .Net, ASP.NET, CLR -- all of it is technology they want to keep.
This. It might appear a bit clunky at first sight, and it might not provide the shiniest eye candy out of the box but in terms of development velocity to get something practical done - provided it is combined with the tooling in Visual Studio - it's really nice.
I agree I would not suggest doing anything non-standard with it. And yes, it looks dated and ugly. But in some places visual appearance of widgets is not that important.
Often, boring UI:s are good. The tool should fade into the background and let the user focus on getting her work done.
While I love shiny things the real world accepts good enough solutions that are economic to develop.
Theoretically I just adore WPF:s scenegraph oriented paradigm since it feels the right way but... I've observed that for a large projects practically it needs a lot of work to get anything usefull done compared to forms. Perhaps the overall architecture is a bit big-org oriented where every tiny widget will have its own development team. Which is understandable but means for simple UI:s WPF might be a lot more expensive.
I haven't done anything massive with it myself but have just observed a few projects from my work. I hope there are counterexamples.
"Non-standard" doesn't necessarily mean "shiny".
Often you need to compose elements in a way that makes a certain UI feature possible, or convenient.
Such compositions aren't inherently shiny; they can even be ugly. But functional.
I love using simple boring UIs because they are predictable. WPF encourages non-standard custom drawn stuff that only serves to distract from usability in my opinion. WPF also still has weird focus rectangles and font issues.
