For the first version of Catalyst, we're focusing on the one-way workflow from Catalyst to Builder (start in Catalyst, finish in Builder). We realize this is a big limitation and hope to address soon after the first release.
There are a number of partial solutions for this problem, which you can find by searching around the forum. Most of them involve dividing the work between Catalyst and Flash Builder as cleanly as possible, which is surprisingly easy to do given the new skinning architecture. Here's an excerpt from a post I wrote that seems to have been lost:
"If you are iterating between Catalyst and Builder on a custom component, you may find it is worth splitting the custom component out in that fashion. I'm planning to write an article explaining this, but here's a preview:
In Builder, create a new subclass of SkinnableComponent. Put the component logic in there.
In Catalyst, make sure that your custom component has those parts and states.
Then import the custom component into Builder.
In the MXML file for the custom component, change the root tag <Group> to <Skin>. It's now a skin (not a custom component).
In the MXML file for the skin, add the [HostComponent("com.me.MyComponent")] metadata (see HScrollbarSkin.mxml, in the Flex 4 SDK, for an example).
Wherever you are creating the instance, change the code <MyComponent/> to <MySkinnableComponent skinClass="MyComponent"/>. Now you're creating an instance of your skinnable component and assigning it the skin created in Catalyst.
Done! If you make more changes in Catalyst, you'll have to repeat steps 5-7 again (for now!)."
I'm relatively new to object oriented programming, but I was wondering if it was possible to create a main program in Flash Builder that overloaded functions in a secondary Flash Catalyst project? If so, then it shouldn't be necessary to modify the FC project directly, thus allowing round tripping...
Is something like this possible?