The target of "Applications" is heavily tied to lots of enhacements like annotation/metadata, DI, AOP and so on... All this kind of "commodities" are the kind of things that could make FlexJS a better option over the miriad of JS frameworks out there. All this things make easy to configure or change the behaviour of your code with much less effort thand classical programming (the tipical example of logging based in an aspect is what I'm referring to) . As well maven (or gradle) support is other important feature to have into account (in this way, I started some days ago a work to mavenize falcon compiler and then I'd like to make the same with FlexJS) and this will make lots of people take into account this software only becasue they can integrate easily in their stablished CI and build process.
2013/7/17 OmPrakash Muppirala <bigosma...@gmail.com> > On Jul 16, 2013 5:14 PM, "Alex Harui" <aha...@adobe.com> wrote: > > > > > > > > On 7/16/13 4:51 PM, "Carlos Rovira" <carlos.rov...@codeoscopic.com> > wrote: > > > > >That would be right if huge applications where created in a consistent > and > > >homogeneous way during all "development life". If the app is tiny > there's > > >more opportunities that experienced programers make a good work and code > > >in > > >the overall. > > > > > >When the app, the team, the code grows over years, more difficult is to > > >see > > >something in that way. For example, we have a product started in Flex 4 > > >alpha, back in 2009. We started with a framework and API that was not > > >finished, we had less experience than we had today, experiment and learn > > >lots of things in this 4/5 years. To get all this knowledge in our app, > we > > >need to invest time to upgrade different pieces at different time. Today > > >we > > >continue evolving the code base, to simplify old code, readapt, to new > > >best > > >practices incorporated, and so on, through all codebase, trying to not > > >break things. This is a huge task not all people and companies want to > > >perform...take into account that this is a project with 160+ projects > (80+ > > >flex), so changes and improvements in all code base could take a year to > > >reach all modules. > > > > > >To rewrite this kind huge product in FlexJS...I think one day we could > get > > >it possible, but that implies we continue evolving and preparing our > app, > > >and converge with FlexJS some time in the future when flexJS would be > more > > >robust and finished, and event there it would be a monumental task since > > >the product still continue growing.... > > > > > >So it's possible...yes, in theory, but I would love to see how many real > > >cases we could get (huge real cases) > > > > Yes, I'm interested in finding folks with huge real cases as well. The > > main question is: to what extent are low-level Flash APIs used in your > > app? I had my internal customer grep their code for "import flash.*". > > Most of their imports were for low-level events. This search will show > > you where you used Dictionary. > > > > The harder-to-answer question is about other Flash APIs that aren't > > supported easily on JS like weak references and e4x. You'd have to > search > > every addEventListener call and find where you used that additional > > parameter. You can search for XMLList and find most places where you > used > > E4x. > > > > Then there is the even harder-to-answer question about Flash APIs that > are > > part of the component API that we got for free from Flash like > > useHandCursor or setting filters and blendModes or shaders, and transform > > manipulations. > > > > But otherwise, to the extent your app is MXML components glued together > > with ActionScript that doesn't rely on these hard-to-mimic things, it > will > > be interesting to push this code through FalconJX and FlexJS and see > where > > it falls down. Lots of APIs from existing Flex components will be > > missing, but I hope to see folks contribute these missing pieces over > time. > > > > I don't know Parsley or Swiz or the other app frameworks, but I'm > > wondering how reliant they were on any of the Flash stuff and if they > will > > just cross-compile or not. It'll be fun to find out someday... > > I am sure there are others in the list with more knowledge about this, but > custom metadata and describeType is probably very important. > > Om > > > > > -Alex > > > -- Carlos Rovira Director de Tecnología M: +34 607 22 60 05 F: +34 912 94 80 80 http://www.codeoscopic.com http://www.directwriter.es http://www.avant2.es
