incubator-flex-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Ben Dalton <bendal...@gmail.com>
Subject Re: Flex 5 in haxe
Date Fri, 16 Nov 2012 21:32:05 GMT
Another couple pennies worth of input...

One of the things that isn't being considered is that even though you can
compile the Haxe language back down to many different bytecodes, it's still
not a way to directly remove the dependency on Adobe's runtime. So much of
the core of Flex is dependent on the flashplayerglobal libraries which
won't be present in other runtimes.

My guess is that in order to achieve any sort of independence, we'll need
to first consider the targeted runtimes and figure out what's the minimum
common denominator feature set we feel comfortable depending on and
rearchitect the framework around that functionality being available to our
framework and abstracted out into a common interface.

I don't think it will simply be sufficient to "rebase" our framework on top
of another runtime and attempt to bridge the gap between the target runtime
and what was offered in FP. That sounds like a recipe for headache at best.

My vote for all of this is to keep dependence on AS3 and FP for Flex 5.
Focusing on improvements in performance and load times while focusing on
compatibility with AIR as a major aspect of our development.

After Flex 5, we can reconsider targeting other platforms or other
languages for development.



On Fri, Nov 16, 2012 at 1:20 PM, Fr├ęderic Cox <coxfrederic@gmail.com> wrote:

> Personally I think we are looking at the wrong problem here. The changes
> proposed seem quite drastic to me. I believe we can achieve great things
> using the current codebase and target. Sure, we need to remove the
> dependency from Flash at some time but we need to be realistic here ..
> Porting too Haxe, porting to HTML/JS, ... they are not realistic targets
> for now due to the limited features compared to what we currently have.
> After that we can look at porting to a target that has the same feature
> capabilities. This can be HTML5 but at this moment this is not the case so
> I believe we shouldn't look at too drastic options which might bring this
> project down. Let's produce results by fixing bugs and improving
> performance. You'll see that a lot of the "frustration" around performance
> is due to the code in the Flex framework and not AS3.
>
> On Fri, Nov 16, 2012 at 10:13 PM, Bruce Montague <
> Bruce_Montague@symantec.com> wrote:
>
> > Hi, I do not know that much about Flex, the history of Flex, or all the
> > other background relating to this thread, etc.. How realistic would it be
> > for an open source community to write an open source, portable, (likely
> > vanilla C), equivalent of  Flash Player and the AVM? Maybe not including
> > all of the features, but a good-enough subset? Haven't there been a
> couple
> > of attempts to do this, some with earlier versions of donated code? (Was
> > that really open source code?) Why did none of these efforts succeed?
> > Patents on codecs and the like?
> >
> > Thanks,
> >
> > -bruce
> >
> >
> > -----Original Message-----
> > From: Gordon Smith [mailto:gosmith@adobe.com]
> > Sent: Friday, November 16, 2012 12:58 PM
> > To: flex-dev@incubator.apache.org
> > Subject: RE: Flex 5 in haxe
> >
> > > From what I previously read, I don't think we were getting an updated
> > Falcon compiler that will generate AVM3 code.
> > > They were not planning on open-sourcing that (but correct me if I'm
> > wrong in that aspect).
> >
> > That's correct. Adobe has no plans to open-source its new AS4-for-V12
> > compiler.
> >
> > - Gordon
> >
> > -----Original Message-----
> > From: Nicholas Kwiatkowski [mailto:nicholas@spoon.as]
> > Sent: Friday, November 16, 2012 12:27 PM
> > To: flex-dev@incubator.apache.org
> > Subject: Re: Flex 5 in haxe
> >
> > On Fri, Nov 16, 2012 at 3:08 PM, Stefan Horochovec <
> > stefan.horochovec@gmail.com> wrote:
> >
> > [snip]
> >
> >
> > > The development of the new VM and AS4 specification is not reported or
> > > discussed with Apache Flex, knowing that we depend exclusively of
> > > Flash Player and AIR to execute applications. This in my opinion is
> > terrible.
> > >
> > > Again we will have to wait for an update from Falcon to generate code
> > > for the V12, this delay the progress of the Flex when we are expecting
> > > more and more code from Adobe.
> > >
> > > I'm not saying that we should use haxe, or some other compiler, just
> > > think the time is an even broader discussion. The Flex should continue
> > > only with Flash Player / Adobe AIR runtime?
> > >
> >
> > This is, and has been par for the course.  When Adobe doesn't want to
> hear
> > us whine and moan, they close off development.   It happened before (and
> in
> > many more products than just Flex/Flash), and I'm sure it will happen
> > often in the future.  They feel they can surround themselves with
> > "stakeholders"
> > (a small, select subset of customers that their marketing team found
> them)
> > to make major changes to platforms, products, etc.   At least this time,
> > they were pretty clear in saying we wouldn't have a seat in the table for
> > the future.  Previous times they gave us the illusion that we did.
> >
> > From what I previously read, I don't think we were getting an updated
> > Falcon compiler that will generate AVM3 code.  They were not planning on
> > open-sourcing that (but correct me if I'm wrong in that aspect).
> >
> > I'm pretty sure the community as a whole over the last 11 months have
> > determined to break our dependency of the Flash Player.  We've had a LOT
> of
> > proposals on how to do it -- none of them executed yet.  The silliest of
> > the bunch in my opinion is porting to HaXe or starting from scratch.  Our
> > power is leveraged from Adobe's initial 200,000 hours of labor over the
> > last many years.  We got an awesome code-base that while, it needs some
> > major tweaks, is is really good shape.  Dumping out the baby with the
> > bath-water is not the way to go on a platform that is mature and used by
> > MANY large enterprises.   That being said -- this is now in the Apache
> > world and I can't stop anybody from doing that, but I won't also be
> > helping redo Flex from scratch.
> >
> > -Nick
> >
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message