Thanks, will do. From what I know about starling, it's just a matter of using a different set of classes for some of the basic display object types. Would doing just that suddenly make Flex just work with it? I don't know enough about Flex to really know that. In terms of the Stage3D stuff though, Christer Kaitila has a good book on the subject that I've read through but I haven't really done a lot of work with directly yet. So I imagine that just getting Flex components to be hardware accelerated via starling would in and of itself make a lot of these mobile performance complaints go away if it was possible to do it. However, if we could get some new components to leverage the stage3D in a data driven way, I think that would certainly be something that could help boost Flex in a positive direction after the transition.
Thanks, David Hertenstein Lead Developer .idea T. 214.529.0668 E. david.hertenst...@idea.com www.idea.com -----Original Message----- From: omup...@gmail.com [mailto:omup...@gmail.com] On Behalf Of Om Sent: Thursday, February 28, 2013 5:06 PM To: users@flex.apache.org Subject: Re: Future of Flex technology On Thu, Feb 28, 2013 at 2:48 PM, Hertenstein, David < hertenstei...@idea365.onmicrosoft.com> wrote: > With Adobe obviously focusing on improving the runtime for the gaming > stuff, that in theory should be good for Flex if the new features can > be utilized by class updates in the framework right? I'm pretty > knowledgeable and familiar with Actionscript as it has matured over > the years, but not as much with the inner workings of the Flex source, > but from my understanding hopefully I'm on the right track there. > > With Flex being focused primarily on data driven apps, as opposed to > custom applications (games) would it be possible to create some new > Flex components that utilize Stage3D? I know there are some simple > controls that have been built using starling, feathers UI, but I > imagine the same technique used there could be applied to Flex > components. Also maybe there could be some Flex components related > more to the 3D context itself. My understanding is that the display > list which is the normal 2D graphic pipeline sits on top of an > underlying 3D context that is initialized but maybe Flex could be useful in > adding some interoperability between the two. > So possibly a set of components that renders out to the 3D stage as > opposed to the display list. That might be a little too abstract of a > description but it would be great if there were not only components to > create stage3D contexts in a Flex application (I think that you are > supposed to be able to have multiple contexts in a single > application), but also to be able to pass in something to a context > via a flex component. So the component would initialize the context > and make sure it's visible and size it etc, but then you could nest > things in the component in order to control the innards of the context itself. > > Just a few thoughts, but these are the kinds of things that I imagine > in terms of Flex maybe being able to utilize new features from the > runtime as they come out, besides the cross compiling work. > > David, you are indeed on the right track. This is effectively the approach I am experimenting with, currently. Although, Stage3D is all a bit new to me and it is taking a while for me to get accustomed to working with those APIs. One thing I realized was that to have any chance of success, we need to re-write a lot of existing Flex code or start from scratch with this approach. Luckily, Alex has started working on his new ASJS framework - essentially re-imagining the Flex framework. I hope to follow him and see how much I can abstract the display list rendering out and use Stage3D instead. If you are subscribed to d...@flex.apache.org, please look out for more emails about this approach in the future. Unfortunately, with the Installer, Mustella CI, OneClickMustella etc. related work, I dont have a lot of bandwidth currently, but I hope to make some time soon. If you are interested, you could come join the fun! Thanks, Om
