I think it's really weird that people point to cancelling the Courier project as an example of a bad decision - all we ever saw from it were some leaked concept videos! Without knowing the actual state the project was in - it might've been a huge mess, as the Kin was - surely it's impossible to have any insight into the wisdom of continuing it.

It sounds like Microsoft is adaptive and pragmatic. What's not to like about that?

If everyone was using Silverlight today I suspect they wouldn't be saying the same thing. Instead it has virtually no traction among devs (aside from a core group of hardcore Microsoft boosters), it has little legitimacy, and the smartphone explosion has rendered it dangerous to choose.

If Flash is a relic of the past (true), then Silverlight is a relic that didn't even have relevance in the past.

  It sounds like Microsoft is adaptive and pragmatic. What's not to like about that?

I really wish this was all there is to it, but their behavior doesn't qualify as adaptive and pragmatic as far as I'm concerned. It qualifies as aimless and schizophrenic.

It's one thing to suggest, fail and then drop technology. Google does that. Even new Apple does that from time to time. It's another thing to suggest, fail and then keep developing technology in some state of semi-cancellation while backing a completely different approach at the same time: for instance, Silverlight. Right now, it seems like a classic case of throwing stuff at walls and seeing what sticks. That's part of the reason why Microsoft is struggling in any area except for its core business.

I don't know that Silverlight is in any state of semi-cancellation, they're just now promoting it mainly for desktop and mobile apps instead of for web apps.

Yes, it is replacing WPF.

Right, but you have to take another step back. As you say, it's good that Microsoft moved away from Silverlight because it didn't gain traction and was wrong for the market.

But before we laud them for moving away from yet another high-profile flop, why didn't it gain traction? Why was it wrong for the market? Because it failed to learn the lessons of the past. It is yet another not-invented-here, one-stack-to-rule-them-all product from Microsoft. It needed developer mindshare and it was going against an entrenched competitor in Flash.

Why was it even begun, amounting to a massive waste of developer time, money, and opportunity cost? That's the real problem at the core of Microsoft. Everybody's a dreamer, everybody's a star, failed projects pad manager resumes on 1 Microsoft Way.

My theory is they wanted some way to get a bit of return on investment from WPF and thus WPF/E which became Silverlight. In 2007 or so who was to know that Apple would have the clout to almost singlehandedly drown RIAs in the bathtub.

Silverlight is really just a pared-down, more portable, and somewhat decruftified version of the .net framework. They probably figured it would be valuable to have such a thing even if the initial "browserplugin" usecase didn't pan out. And in fact it has been.

I would say resources to learning silverlight would be a big reason it hasn't gained traction. After the basics it's a real struggle to work out how to do a heap of things, the answers usually seem to be buried in blogs and comments if at all.

I find MSDN equally hard to work out for other technologies as well, but at least there are other resources out there, with silverlight sometimes the only mention of a feature existing is on a developer blog.