>In the spirit of understanding, with a prototype we are going to develop it in >a non intrusive way to the existing compilers. There is no plan to go hacking >into the existing code if we don't >have to at this point, plus that would be >completely stupid IMHO. > >I think the proof of concept will use the FlaconJx and ASDoc compilers to >prove our configuration concept, which verifies we will not touch the existing >MXMLC or COMPC code for now. >
I am all about this. I already started porting all of Randori to AS and plan to check it in here in my whiteboard Right now we can't compile it with Falcon in a way that will run, but it is an excellent gap analysis of what I need to help write. Randori is about cross-compiled business logic, integrated with existing JS components and intrinsic Dependency Injection. We are deploying apps this way today but using C# as our source language. The apps work, and the approach works well. Ironically, the most difficult part is finding reasonable JS components that aren't full of coupling, but let's hope that these continue to evolve or some of the efforts here on Flex provide us with an excellent JS component set going forward. In either case, this is where I plan to spend my time, Mike
