Just to be clear -- there is no AS4. Adobe proposed the standard and started some work on it and then didn't proceed with it any further.
I don't think we need to set a goal of a complete re-write to read a 5.0 milestone. I think working on some of the bigger concepts of decoupling some of the components to make them ready or at least more portable to FlexJS is a goal we can set. Setting the bar too high I think will make it unattainable. -Nick On Tue, Dec 9, 2014 at 3:43 PM, <f...@dfguy.us> wrote: > I think it's a touchy subject. The flash player and adobe always seem to > get a bunch of flack, often unlike most other similar platforms. This whole > "proprietary" phrase is usually used as some sort of bludgeon meaning > intrinsically bad; which it obviously is not. But it's a difficult > situation though like mentioned where you have to depend on a company for > support and there's a lot of uncertainty there. > > So either rewrite the platform to use a new AS4 or rewrite to use a > different runtime like just JavaScript in the browser, you're still doing a > lot of rewriting. Then, the browser isn't exactly a runtime in terms of > being interpretive as opposed to precompiled. So if you target the browser > you still have to figure out some "third party" framework for mobile > deployment in app form right? > > What I'd like to see is the ability to something like JavaScript as an > option but then to actually continue to have a super charged runtime to > target that's actually supported to the fullest. I don't know if it's > really Adobe's fault regarding the plight of flash player or if it really > just boils down to outside forces screwing things up in one way or another. > > If flex became a JavaScript only framework for targeting would that > actually help it get used more, or would there be even more competition. I > think what might be better would to be to have a strong runtime supported > by a "proprietary" company properly, and then also enable to use of open > standards, html5, JavaScript etc within the framework itself so that you > could reuse a lot of the web standards code that's out there. > > David > > > > -----Original Message----- > From: Jesse Nicholson <ascensionsyst...@gmail.com> > To: dev@flex.apache.org > Sent: Tue, 09 Dec 2014 1:44 PM > Subject: Re: Let's talk about Flex 5 > > My 2 cents is that the project should be focusing on moving away from a > third-party, proprietary and frankly dead platform. I'm also not sure what > actionscript 4 has to do with anything, it's not like there will be a > ground-up rewrite of the entire project to port it to a new language. Even > a mass, automated conversion (were it possible) would be, in purpose, > purely for the benefit of yet again, maintaining compatability with a > third-party, proprietary platform. In fact it would be worse than that, it > would bind the project specifically to that third-party, closed source > target runtime. It's discussions like this that are at the root of the > previous concerns I've raised which can be summarized as such: is flex it's > own product, or is everyone just working for adobe. > > I know it's a lot of work, but I really think we need to replace AIR and > Flash Player as the target for this project. Doing so would create a truly > standalone, apache driven product future-proof, omni-platform application > development. I didn't want to really get into this discussion yet but, > since it came up... > > As for the question of AIR on iOS, Adobe created a LLVM frontend to convert > actionscript to LLVM-IR, which adds all of the benefits of LLVM for > optimization, then emits ARM. The runtime (which is probably C++ source, > since the entire Tamarin engine is C++) is precompiled already to AMR libs, > linkin is done, etc etc, lots of proprietary magic and you've got an final, > native assembly. This is called AOT or ahead of time compilation, rather > than using JIT (as the flash player uses) in order to comply with anti-jit > license terms of IOS development. > > The AVM2 is (mostly) open source under the Mozilla Public License, but has > recently gone missing from mozillas mercurial repos. It's tough to track > down, but you can still find "tamarin-redux" and get a zip of the source > tree. I believe the AOT compiler source code is present as well. > > On Tue, Dec 9, 2014 at 2:07 PM, Gary Yang <flashflex...@gmail.com> wrote: > > > just to bring this up... > > > > On Fri, Jun 6, 2014 at 1:45 PM, Gary Yang <flashflex...@gmail.com> > wrote: > > > > > Java Spring is a very good example for framework evolving, for Flex, I > > > think it is the similar situation: > > > > > > The key function is > > > 1) Mxml > > > 2)Binding > > > 3)Data structure such as IList implementations > > > > > > on top of Mxml/Binding/Data is 1) UIComponent 2) Skinning 3)network > > > components > > > > > > and then osmf, reporting, text .... > > > > > > I think from Flex5, we should modularize these into different kinds of > > > projects: > > > > > > 1) Flex data, to provide an infrastructure for 1) mxml -> as > generation, > > > 2) reactive programming(Binding); 3) related data structures; > > > > > > 2) Based on Flex data, Flex UI, to provide the basic UI implementation( > > in > > > Flex4 skinning way ), could be multi projects. > > > > > > 3) Based on Flex data, Network/Native components, http, websocket and > > peer > > > to peer, native devices communications. > > > > > > and then projects that specified in different fields: video, text, > > > reporting .... > > > > > > I see Flex as a tool sets to solve complex user interface, so Flash > > player > > > is the only way to work in desktop web, so the language has to express > at > > > least as much as AS3, so if possible, using Java as a language will be > > > enough to downgrade into most other languages. > > > > > > Just a little thought. > > > > > > -Gary > > > > > > > > > > > > > > > > > > On Thu, Jun 5, 2014 at 8:06 AM, <f...@dfguy.us> wrote: > > > > > >> Right, it could be written in it and not cross compiled. It could be > > that > > >> the project is what get cross compiled and then packaged with the > > runtime. > > >> I think though that this could be a good opportunity to improve the > > runtime > > >> in general as was previously talked about with the AS4 plans but I > guess > > >> we'll have to wait and continue to ask Adobe to work on it. > > >> > > >> David > > >> > > >> > > >> > > >> -----Original Message----- > > >> From: Harbs <harbs.li...@gmail.com> > > >> To: dev@flex.apache.org > > >> Sent: Thu, 05 Jun 2014 6:56 AM > > >> Subject: Re: Let's talk about Flex 5 > > >> > > >> I don’t think that’s correct. Unless I’m mistaken, the AIR iOS runtime > > is > > >> written in Objective C from the get-go. (Although it might be written > in > > >> C++. Dunno…) > > >> > > >> Whether or not it makes sense to rewrite the AIR iOS runtime in Swift > is > > >> an entirely different question — which probably only the engineers at > > Adobe > > >> could really answer… > > >> > > >> On Jun 5, 2014, at 2:51 PM, f...@dfguy.us wrote: > > >> > > >> > My understanding is that the entire runtime gets cross compiled into > > >> objective c. So Adobe would have to rewrite this to use swift, but I > > think > > >> the same process would basically be used. It's possible though that > > swift > > >> could enable additional features. The limitation on loading compiled > > byte > > >> code is purely just a licensing and not a technical limitation that > > imposed > > >> by Apple. > > >> > > > >> > David > > >> > > > >> > > > >> > > > >> > -----Original Message----- > > >> > From: Harbs <harbs.li...@gmail.com> > > >> > To: dev@flex.apache.org > > >> > Sent: Thu, 05 Jun 2014 1:17 AM > > >> > Subject: Re: Let's talk about Flex 5 > > >> > > > >> > Really? The only way I know of outputting ActionScript for iOS is > > using > > >> AIR for iOS which is just a swf with an embedded runtime. > > >> > > > >> > On Jun 5, 2014, at 4:26 AM, f...@dfguy.us wrote: > > >> > > > >> >> There's already the ability to cross compile to objective c for iOS > > >> > > > >> > > >> > > >> > > > > > > > > > -- > Jesse Nicholson >
