1 Reply Latest reply on Nov 21, 2011 12:02 AM by ccapndave

    Future of large Flex applications


      So - we have invested heavily in Adobe Flex as our client application tier (I bet no one else has invested heavily in Flex on this forum )...  I have a few questions for the community, which hopefully can be discussed in terms of what is going to make the most business/technical sense for development teams going forward and not get into religious debates...


      Here are my two questions:


      1. Does anyone know of a good framework/IDE/debugger for application development that can run across platforms/screens (e.g. browser, desktop, mobile), is efficient for developers to use, and provides rich visualization components than can be easily customized?  Even if not multi-platform/screen, what is something that would give us a Mate/Swiz-like (IoC) client application development workflow (with binding, good cross-browser skinning, etc) that doesn't depend on the flash player (e.g. runs inside the browser w/o a plug-in)?
      2. Given that there is some option (or one that will mature over the next year to be viable), what migration path should we consider for our 250k lines of flex code?


      Those might be premature questions, but we have delivery teams in our organization starting to investigate other technologies for the next generation of our client application tier (which prior to the last two weeks was nothing anyone had mentioned).  I want to make sure we get ahead of delivery teams in our org starting to use different technologies to deliver projects, which would likely end up with massive technology/framework fragmentation and a maintenance nightmare.  Basically - I need to now revise our technical roadmap for the client tier of our application suite and get our teams and execs to rally behind it, based on the recent happenings regarding Flash and Flex.


      To provide a bit more context...  We have upwards of 250k lines of code in our flex projects, all pulled together into two main web applications via the use of modules and runtime loading.  Most of the applications are deployed in the browser.  Some are deployed on the desktop via AIR.  Some are deployed on mobile devices via AIR. All share a material amount of reusable Flex code that are in libraries. We use Mate and Swiz for our frameworks on top of Flex.  We use degrafa for skinning for our flex 3 apps and for some projects - flex 4 spark components.  We have complex visualizations (using just about everything flex and third party libs have to offer) and complex data grids (using ADGs and customized ADGs).



        • 1. Re: Future of large Flex applications

          Sproutcore (http://www.sproutcore.com/) is an open source Flex-like framework for HTML/CSS/JS with MVC, binding and an ORM layer.  Its a very impressive bit of work, but be warned that it is nothing like as powerful or mature as Flex.  Also in my experience Sproutcore compared to Flex takes significantly longer in development and maintenance for an equivalent or similar product.  This is mainly due to lack of decent tooling (lack of IDE error checking, except in simple cases of syntax) and the need to test and maintain across multiple browsers and platforms.