Well, I got the feeling that some users on this list were advocating a total
rewrite asap rather than maintaining and improving the current codebase. A new
Flex framework, built in Haxe or some other language than AS3/MXML is a totally
new framework that will have no support for the current codebase, which means
I'm either stuck with the current technology or I have to rewrite all my stuff
and unfortunately rewrites just-for-the-fun-of-it are hard to sell to clients.
I agree that the strategy should be maintain and enhance the current framework
while planning/preparing for the future.
Yes, unfortunately Steve Jobs managed to kill Flash in mobile browsers (ok,
other things contributed as well :-), so I can't deploy to Flash player in
mobile browsers, but AIR is a perfectly usable solution to that problem so I
can easily deploy to an app for Android/iOS with the same code base.
"writing yet another Gui framework on top of HTML/JS/CSS"
Noone is proposing such a thing.
Flex needs to be cross platform and with OOP language.
I didn't mean that a new Flex would be written IN Html/Js/Css. But some see
the solution to the Adobe VM dependency being to deploy to Html/Js/Css,
generated by Haxe or some other tool. I don't, not because there is anything
wrong with Haxe or whatever other tool would be used, but simply because HTML5
still has years to go before it can support the data-rich apps that Flex in
Flash player/AIR excels at, and that can't be fixed at the compiler level
because in the end you just get Html/Js/Css that the browser executes, with all
the plusses and minuses that come with that technology stack.
But the question is: does keeping AS3 make it possible to target OTHER runtimes
like HTML5, or any other new in the future?
I guess the answer to that is probably no.
Another question though would be do you want to target HTML5 as a runtime ? I
certainly don't for now at least and many small miracles have to happen in the
HTML/browser world for that to change for me. Keep in mind that most mobile
development development today fex is via native apps. This is because the
HTML5 browser apps can't compete either on the development side or on the
execution/functionality side.
"Which means that there is plenty of time to wait for other, better deployment
possibilities to come along"
I don't understand what and from who you are waiting somthing new for such a
purpose?
Well, I guess I don't know either what it would be or where it would come from
:-)
I just want to deploy to desktop browsers and mobile apps, preferably with the
same or mostly same source code. If smth other than AS3/Adobe AVM allows me to
do that with the same ease and power, AND it has less risks attached than
sticking with AS3/Adobe VM, then I'd probably use that.
On 16.11.2012, at 14:34, sébastien Paturel wrote:
i don't see any conflit here.
it's timeframe consideration.
In short term we maintain and make little enhancements (like better
performances, with starling for example) to the current framework which has
still viability for several years (not over 5 i think)
And we prepare the long term, with a new code base rewrite.
If you develop browser based front-ends you already has an issue with flex, as
it will not run in tablets browsers !
"writing yet another Gui framework on top of HTML/JS/CSS"
Noone is proposing such a thing.
Flex needs to be cross platform and with OOP language.
"The runtime is free, easy to get and works the same in every browser almost since
time began"
Changing the source language does not mean that we don't deploy to the flash
runtimes.
Haxe is doing that very well.
But the question is: does keeping AS3 make it possible to target OTHER runtimes
like HTML5, or any other new in the future?
"Which means that there is plenty of time to wait for other, better deployment
possibilities to come along"
I don't understand what and from who you are waiting somthing new for such a
purpose?
Le 16/11/2012 15:07, Hordur Thordarson a écrit :
Good points made by Ben and I totally agree that the rewrite-or-not decision
cannot be made without first knowing if this project is to be about maintaining
and improving the current Flex framework, or about creating a new and then
possibly very different framework with little or no legacy support.
I develop browser based database front-ends and I have to say that there is
nothing out there that comes anywhere close to rivaling the maturity,
abilities, development speed and deployment conveniance of the current
Flex/Flash player combo. So while I understand the problem of being tied to
the AS3 VM and Flash player, unfortunately this same combo is a big part of
what makes Flex attractive to enterprize developers. The runtime is free, easy
to get and works the same in every browser almost since time began. Building
database front-ends and data vizualisation tools in HTML/JS/CSS requires you to
give up a large chunk of your sanity, because the technology is immature and
non-standardized, the toolset is years behind what Flex has and at deployment
time you have to deal with endless browser diffs and bugs.
So I certainly had hoped that Apache Flex would be about improving the current
framework rather than ditching it and writing yet another Gui framework on top
of HTML/JS/CSS.
My opinion is, having read what there is to be had, about Adobe's intentions
with the Flash player and ActionScript, that Flash player isn't going anywhere.
And while it is there, it will most likely include the AS3 VM just like it has
had the AS2 VM for years. Which means that there is plenty of time to wait for
other, better deployment possibilities to come along.
On 16.11.2012, at 13:55, Joan Llenas Masó wrote:
Flex to me is:
1. Cross browser/platform
2. MXML.
3. Modularity
4. Spark architecture
I would agree with this, BUT, and for many of us I think this is a big BUT, I
would add #5 which is the extremely easy, well known deployment of client apps
to any desktop where Flash player is available.
Those are my two cents :)
Hörður Þórðarson
Lausn
Iceland
On 16.11.2012, at 13:28, Ben Dalton wrote:
It depends on the goals of the Flex project. If we are looking to create
something that is purpose built for the Flash runtime, then I don't believe
so.
However, if we are looking at Flex as a toolkit to build rich interfaces
and are willing to break from the SWF compile target being our core focus,
then I believe yes.
I think that question really needs to be answered first.
While I love Flex as-is, we are seeing Adobe's priorities change for the
runtime and even some version fragmentation that we hadn't before (pepper
flash player in chrome and the MS flash player built into metro IE). It
pains me to say this but it looks like the standard HTML js CSS world is
receiving all of the developer efforts right now and I think that we should
make that a focus if we want this project to survive and thrive beyond the
next year or two.
There is a conflict of interest here which I'm sure we are all aware of.
The goals of building the best toolkit for rich applications for mass
consumption and the goals of enhancing the current toolkit to continue
servicing our enterprise clients seem to be in conflict.
This is my two cents of an opinion and I will be glad if I am mistaken.
On Friday, November 16, 2012, sébastien Paturel wrote:
After several discussions about the difficulty to break UIComponent for
example, and the whole re-architecturing of flex SDK issue.
I'd like to launch this poll to see if there's a consensus about it.
So in your opinion flex SDk:
1- Need a complete re write
2- Can be re architectured from the current code base
Thanks
