> > > That being said, wouldn't it be advantageous to think about APIs that > could stand the test of major refactors and start migrating or add > extensions to the current code base that would give developers the option > to start moving towards the next framework? > > I mean, a public API is just that, what goes on under the surface is for > the low level developers to think about and implement. > > So you have Flex 5 that starts to really use this new API that maybe uses > some adapters to fit some popular components under the API while creating > more modular revisions at the same time. > > I think some type of simultaneous refactor is possible giving the amount > of brains around here. > > Mike > > -- > Michael Schmalle - Teoti Graphix, LLC > http://www.teotigraphix.com > http://blog.teotigraphix.com > > While that may work, the differences so far described in AS4 and the new VM sound so fundamentally different from what we have in AS3/AVM2 that I don't actually believe this would be possible. Based on rumors not even Starling is compatible, and the display list will pretty much be gone.
I think it would probably be a complete waste of time to do that, but of course I may be wrong. I think the only way forward is to start with a brand new unit test suite, and build a new version of the Flex framework from the ground up, the right way. Modular and unit tested, such that it can grow from a small baby to a full grown application framework as we know it today. Trying to shoehorn the current base so it could possibly take advantage of some features of a new VM seems like a waste of effort. That said, anyone here is free to do what they want really. If it turns out well it'll likely be voted into a release. :) -omar
