Of course, it takes quite a while to read through the documentation for the INSANE number of classes in Flex (seriously, the list takes up 3 huge posters of 9pt text), but once you have a pretty good grasp of what it can do it's not so bad to work with.
Forget rich UI, I'm talking about being able to build a beautiful, consistent, responsive, and fast web application using 2 developers. The same app in HTML/JS would take a football team of developers and wouldn't even come close.
Yes, there are things that suck about Flex as with everything else in the world, but for what you can achieve with it, there really isn't anything that compares. :)
I'm always surprised these type of frameworks haven't caught on more. I built an admin system using Mozilla XUL a few years back. It's similar to flex in that it provides a cohesive framework with layouts and widgets geared to more traditional apps. It was a breeze to develop compared to the normal battle I have trying to get html and js to do what I want. Unfortunately Mozilla have only really got behind it as a way to develop FireFox plugins and seem to have missed the opportunity to promote it as general way to develop web apps.
I've always wondered how much flex is really tied to flash. Now flex has been thrown to the community I'd like to see someone try to modify it to use html canvas, svg or webgl as it's renderer (instead of flash). The low level rendering primitives would probably map quite closely. The bulk of flex is written in actionscript and mxml which could be compiled to javascript and then run directly in the browser.
They did. Flex dev tools are quite good, and they were continually improved since Flex 2.
On the other - I'm glad they at least do this when abandoning the project. How many times do we hear "if they discontinue this, they might at least open the source so we can continue porting"? As long as we agree it's pretty much EOLing rather than open-sourcing, we can be glad that we don't get left with binary blobs noone can reuse in any way.