>I'm looking for some amunition for a rather political battle I face. What are some experiences others have had when deciding in their organization rather to roll their own framework vs. purchases or downloading one? What process have you followed to see if existing frameworks would satisfy your needs?

It's all the matter of budget. With any framework, it may take between two and four weeks for the programmers to become comfortable and productive in it. For a homegrown framework, it takes between one and three programmers to build it for several months before they'd be able to produce anything in it.

Also, consider the differences between lightweight frameworks (just a few classes to build forms on, handle the data etc) versus heavyweights (with their own set of metadata, like dbcX, builders and wizards).

I've tried it all - built a (FPD 2.6) heavy framework, built a light one (just a few classes and a few builders as they were needed), used VPM, VFE, WWWC and two homegrown ones (where I added my own builders and even one wizard, as needed). Personally, I'd go with a lightweight (maybe one of them freebie ones) and then adjust it to my needs as I go. Of course, my opinion is based on the utilities that I already have written, and I'd use them again in/against/over any framework that I'd be using.