btami: I would argue that Sprox is more powerful than FA in every way. Sprox allows for a more declarative representation, and leverages TW to do the dirty work of creating HTML. This makes it more flexible, and adaptable to existing widget technology. Sprox also has a better test suite, and better support within TG. Some history: I looked at the FA code early on, and it had some good ideas, but it was headed in a much different direction. For instance, FA is _very_ coupled to SA, and Sprox is not. FA does their own HTML generation, Sprox utilizes a widgeting library to do so. FA does have good support within the Pylons community, but there is no reason you cannot use Sprox within Pylons… I guess a new blog post on that is forthcomming.

Some history: I asked the team members if they would like to merge when I was working on DBSprockets, and was turned away. So, I guess you can say the two projects began at roughly the same time, and I tried to merge with them, but eventually had to just do my own thing.

Karl G: Thanks for checking it out!

Abstractor: In fact, Sprox is not tied to a database layer at all. You should be able to feed it objects and as long as you have created a provider and widget/validator selector for those class of objects, it will work. This is why we are able to handle both mongo and sqlalchemy without any additional code in the core.

@btami, formalchemy is another implementation of the same concept (take your SA classes and move them to forms)

Thanks for finally doing a writeup on this. This is one area where TG has really improved, and I’m glad you finally decided to “toot your own horn” a little. Catwalk was OK, but pretty buggy, and not very flexible. I found DBSprockets pretty impressive, but Sprox just blopws everything else away. And the new TG Admin built on top of it is top notch. It leaves all other admin/scaffolding/whatever stuff in the dust. Just my opinion, of course, but…well, its just awesome.