After last days news I was digging into Starling and cover lot of stuff to get some initial knowledge. I really love what I see. I need to continue the research more things like StarlingMVC, feathersui, and so on...
IMO, the ways we have ahead to choose are: a) Refactor actual Flex framework to rely on Starling. This will prepare Flex to the AVNNext and AS4. This seems very laborious for what people says here, but this refactor will open the door to old applications target the new AVMNext. Problems: We need to know more about AVMNext and AS4, what about localization? accesibility?... until be don't know all this things we can't evaluate this option in its totality and see if it's worth the investment. b) Make a new light weight flex framework *from scratch* that will be a dream platform to develop with OOP, a new evolutioned compiler (Falcon) and lots of new development facilities and be prepared to target new AVMNext and HTML5. This will have in mind composition over inheritance, modularity, and decoupling of libraries. We would need to target as many features of the actual flex framework as possible: * Components (This could rely on feathersui, if the evaluation of this component set will be positive. I already need to see more about this project). * Containers * View States and Transitions * Binding * Application framework with Managers (layout, system, popup,...) and all the things to avoid us to seek for external components and scripts that integrates all the parts. * Styling through CSS (seems feathers have already skinning) * Skinning * Declarative programing through something like MXML * Effects * ... We should take care that this new framework could target multiple runtimes (Flash and HTML5). Without this...we could not expect to have the needed traction this project requires. We could start with Flash in the beginning, but making the needed room to hold HTML5 in the near future. Again the main problem here is to know more about new AVMNext and AS4 to see if this is possible (localization? accesibility? ...), but as well we would want to know more like... there are generics? metadata evolution? ... to take into account in the new frameworks design. To end: This project has many possibilities and the ability to be the platform of choice for lots of people that wants to develop serious RIAs, but as well for people that wants to create complex games. -- Carlos Rovira Director de Tecnología M: +34 607 22 60 05 F: +34 912 35 57 77 http://www.codeoscopic.com http://www.directwriter.es http://www.avant2.es
