I'm a bit disappointed in myself. I haven't been keeping up with Flex 2.0 the way I should have, and it really looks pretty awesome.

I spent some time with one of our talented programmers today, refactoring a form validation framework for Flex 1.5. Flex's original validation model was well-intentioned, but (truth be told) somewhat poorly implemented, and didn't lend itself well to server-side validation. We require server-side validation because we want to write our validation code once, and once only.

So luckily, between Russell's in-depth knowledge of the quirks of Flex 1.5's validation model, his elbow grease, and my pontification by way of UML sequence diagrams, we were able to refactor out a large chunk of a validation framework that will integrate beautifully with Flex's front end validation features while running all the validation logic server-side. The validation logic will then be accessible for other views and for the business tier of our applications (where necessary), allowing us to achieve true reuse. The best part is that the framework is practically 100% transparent to the developer, because the mechanics of the framework are hidden away in genericized utility classes and entities.

I'm pretty excited to see where it goes. If it's as good in practice as it looks on paper, we'll give it to the community.

That being said, I'd be equally happy to see new validation features in Flex 2.0 completely eclipse our need for such a framework. It's obvious to anybody who watches the product demos that Macromedia Adobe really paid attention to the challenges of Flex development, and came up with solid time-saving solutions. I watched a video on the States and Transitions this evening (which are now all completely declarative) and marveled at how much time and ActionScript I could have saved myself, if only I had access to those features in Flex 1.5.

Time to download the beta.