Reflex News

[Ben Stucki]

Hi All,

Some big Reflex changes have been going on in preparation for 360|Flex DC. Most of them are cool and technical (and Tyler's saving those for his presentation), but I wanted to give you a heads up on a big one before the show.

I recently decided that after this 360|Flex I'd like to split from the project as it exists and work on a component set solo. Normally this would involve a new name and lists and build servers etc., but luckily for me (and after some begging) Tyler was cool enough to let me stick with the Reflex name and will be taking on the burden of branding a new project. It was a hard decision for me to give up on collaborating, but I'm excited to see the progress of both projects and look forward to even better results through some friendly competition.

So, this list and the websites associated with Reflex will focus on my continued progress, and I'll be notifying you of any major updates and releases from the competition (that's Tyler) as more details emerge after 360|Flex DC. It should be interesting, so stay tuned.

- Ben

[Patrick Lemiuex]

Yo Ben,

will there be a video of this, I want to move to reflex! I'm most
interested in this project!

P

[Ben Stucki]

Yo Patrick,

I'm not sure if they'll have official recordings, but I'll make it a point to record Tyler's preso. It looks like he's already making some progress on the competing version so it may deviate in some details from the "Reflex" main line, but either way you'll be able to check it out. I'll also work on some getting started content for the main line soon. Thanks.

- Ben

[Tyler Wright]

Reflex has been a good home for building components this last year but I know Ben will take good care of the project as we move onto something new. This is an exciting progression for everyone as it results in even more innovation for the community to look forward to. As developers, components are central to our workflow and often our source of joy or misery in our projects. Particularly as mobile patterns and approaches redefine our UI's it will be essential to have better platforms and greater tools. These are the initiatives we have all been working towards. I'd like to introduce the next step in this progression.

Flight Stealth has been a project once before and it is again. The last post, to a mailing list of only two, was in May 2009. This is the first post of the new project, with many more to come. Flight Stealth will be hosted under the umbrella of Flight XD where advanced binding and non-Flex MXML workflows were formed. I may post a couple of follow-up emails to point out important updates for this new project, but after next week project communication will be on the flight-...@googlegroups.com mailing list. Also be sure to catch the 360|Flex presentation "A Component Set Once Called Reflex"  :)

Each of these projects are open source, representing a significant investment by their contributors, and are fully for your benefit. I encourage everyone to be supportive of advancement and progression in whatever form it takes over the coming months and wish Ben the best in his development initiatives. And thank you for your invaluable contributions.

Tyler Wright

[Jacob Wright]

The Google Group is here: http://groups.google.com/group/flight-stealth for those who want a quick link for joining.

Jacob

[elffikk]

hi,

while I am happy about the progress there are a few things broken
like SlideBehavior and StepBehavior due to Position refactoring
also if I change ScrollBar position the UI is not changed, the reverse
is ok - if thumb is dragged position is changed

I need that fix quickly, doing my own behaviors or fixing yours will
take more time then if you will fix
I prefer to work now on skins

thank you in advance

[elffikk]

hi to all,
please tell me that I am wrong, but I had the feeling that reflex will
not be mature and this is a warning for those that will want to spend
their time to learn it

I am looking at frameworks, some developed by you guys (those that
are contributing to reflex), awesome frameworks, but they are all
unfinished jobs...

openflux(http://code.google.com/p/openflux/) - seems that it was just
an exercise before developing reflex, status - beta1, last update -
May 17, 2009
axiis (http://code.google.com/p/axiis/, http://axiis.org/index.html) -
in beta 1 more than a year
flight (http://flightxd.com/) - not reached version 1 release, and Ben
refactored reflex so that is no more needed for reflex (anyhow, thank
you flight for lightweight flex independent binding !!!)
flare (http://flare.prefuse.org/) - alpha version, last release -
January 2009!
degrafa(http://www.degrafa.org/) - in beta 3.+ since 2009

I know that some of them are used by a lot of people in a lot of
projects (just to name degrafa),
but beta status is scaring me, also that means no support, no
documentation and very little community activity,

please don't develop reflex just because you have some sessions at 360|
Flex Conference

[Andrew Westberg]

I hear you. I have projects stuck on Flex 3.x because of Degrafa. I
helped contribute some code to the Axiis project, but I think it's
suffering the same fate due to Degrafa. While we're at it, let's add
AudioSpike to the opensource graveyard list. Awesome demos at 360|
Flex Indy last year, but the audiospike.org website seems to be gone
now.

-Andrew

[Ben Stucki]

You're welcome?

- Ben Stucki

[Tyler Wright]

It's hard to get an OS project to that point of critical mass where it can be sustained long term. It usually starts with a developer or a group of developers who have a common vision. They put countless hours into the project, sometimes with little gratitude, they make several releases, and then a review of the project shows that future work outweighs any future benefit. Either the need wasn't felt, there wasn't much contribution to sustain or there wasn't enough profit to justify the hours and headache. The free code stops flowing because there is more drag then reward. I was on the Degrafa list - the only emails that came through were support requests. I wonder if our community struggles more than others with OS and why.

Please hold off final judgement on these projects. I believe that there is a greater need now than ever for custom components. I hope to have better UIs and better workflows building them, and if not these projects then I guess we wait and hope Adobe responds better to our needs than our fellows.

Tyler

[Tom Ortega]

Aside from what's "not" in the projects, there's something that Tyler sorta hints at in his response that I think is very important to remember.

When Adobe goes to add features, I'd like to think that these open source projects help guide them on what's important to the community. Hopefully the ideas and concepts these projects represent help influence them when they code similar features.

Another thing is to think of these projects as free starter code. I know in the instance of flare, I used it to help me figure out how to do some things. These are very complex ideas being created and given away for free. For that, we should be grateful on many levels. 

Of course, I think the whole point of OS projects is that when you find something missing or an update hasn't happened, then you're supposed to step up and contribute. I approached Stucki in regards to helping with Reflex earlier this year, but then the project I was working on went in another direction so I sorta bailed on him. Sadly, I bet that happens to OS projects alot. However, had I been using it in a project and found bugs that held me up, I would've fixed it and tried my hardest to get that back into the project codebase.

I guess my main point is that we should probably be asking these people, "How can I help?" vs throwing a "You guys suck and take too long."

Thanks,
Tom Ortega
cell: 951.212.9686 | twitter/skype: lordbron
lordbron.wordpress.com area-161.com

[Andrew Westberg]

Tom et. al.,

My comments weren't intended to convey a lack of appreciation for the open source community or all the hard work that goes into it.  I appreciate it greatly and couldn't do half the projects I pull off without it.  I do think that the flash/flex community has some issues with open source when the average project doesn't live beyond a year or two.

From my perspective (taken with many grains of salt), here are a number of possible reasons why open source in the flash/flex community may be having trouble.

1.)  Projects go public too early.  Sometimes there is a project out there publicly before a line of code has been written.  If projects stayed "skunkworks" longer until they could be immediately useful to a developer, this would give them a bigger launch, bigger following, and potentially more people to contribute.  For example, I'm working on an Actionscript library right now that can publish/subscribe directly to ActiveMQ/JMS using the OpenWire protocol for messaging.  It may or may not see the light of day as an OS project.  The benefits to making it open-source would have to weighed against the costs of ongoing maintenance and support by my employer.

2.)  Many successful open-source projects start out with a for-profit company or university program backing them.  The freelancers that make up many flash/flex OS projects are extremely talented, but when it comes to donating time/effort vs. feeding a family, it's easy to see which one wins out.  I feel confident building on top of an OS effort like GraniteDS because there is a company behind it.  I know it's not going away anytime soon because that company's success is tied to the success of Granite.

3.)  Related to #2 is that there are too many paying jobs out there competing for time. Until there are physically more flash/flex developers, this will be tough to solve.  If you blog about flex, have some experience with frameworks and architecture, have a decent following, and let people know you're available, you'll pretty much be turning down work.

4.)  Adobe tends to stomp on air hoses.  For example, the Degrafa fills broke in the flex 4 release.  This simply SHOULD NOT HAVE HAPPENED since the FXG team was supposedly working with the Degrafa team during the design phase.  This seemed to come as a surprise to the Degrafa team as well since I distinctly remember them talking about how FXG wouldn't wouldn't replace Degrafa at 360|Flex Atlanta.  This air-hose stomping caused a ripple effect for down-stream projects like Axiis.  Now we're stuck using FXG which is faster, yes, but not nearly as powerful.  Try doing something like this in FXG.  It's simply not possible without the power of Degrafa GeometryRepeaters.  http://www.flexjunk.com/examples/SmithChartExample/SmithChartExample.html

5.)  Enterprise software development hasn't yet fully embraced Flash/Flex.  Once this happens, you'll have a larger pool of developers who get bored working on something like an ERP system all day and need a creative outlet outside of work.  I think the Java OS community benefits greatly from this aspect.

Anyway, those are just some of my thoughts.  Apologies if my comments/frustrations from earlier offended anyone.

-Andrew

[John Lindquist]

On the flip-side, many open-source projects are successful because the authors use the project in their every day development lives (that's why there's so many open-source application frameworks and tweening libraries). That's also why something like Papervision3D can be a smashing success one day then see the devs be lured away by the Unity3D platform and claim "we just haven't had time to work on it" (which can be frustrating to the users waiting for updates).

If a project lacks that personal motivation, then it needs sponsors for monetary motivation.

[Vasile Ermicioi]

I feel about reflex what Andrew said about degrafa: - "I distinctly remember them talking about how FXG wouldn't wouldn't replace Degrafa at 360|Flex Atlanta" 

and now the development of degrafa if is not stopped at least is slowed down significantly

I had hopes about reflex:

"Introducing Reflex" - "We hope to provide a compelling alternative to Flex that is smaller, faster, and easier to build on top of."

"What project will replace the Flex framework?"

Also there was a promise to have a component set at least as large as the one from Flex 3,

collaboration with FDT team

discussions with degrafa team to have a release for reflex

RPC services

do you remember such things? or may be I read too much and remember too good? :)

now I am working on  a few components,

mostly on skinning with degrafa, 

if I will complete  today or tomorrow a few more components I can share them with  you to use in your sessions at upcoming conference(s)

[Greg Jastrab]

This comment astounds me...

Why are you complaining about someone starting an interesting project.  The Reflex developers didn't sign any contract when they started the project promising to deliver features x/y/z to a 1.0 release by a certain date.  If someone wants to spike on code to present something interesting at a conference they have the right to do so.

As Tom said, contribute instead of complaining.

On Thu, Sep 16, 2010 at 3:59 PM, elffikk <elf...@gmail.com> wrote:

...

please don't develop reflex just because you have some sessions at 360|
Flex Conference

--
Greg Jastrab
Phone: 443.451.7057
Fax: 443.451.7059
SmartLogic Solutions, LLC
http://www.smartlogicsolutions.com/

[Patrick Lemiuex]

I am all for Ben and Tyler's reflex efforts. I appreciate their hard
work and the shit is awesome, so please recognize.

I only wish I had time to make components on the platform.

Thanks,
Patrick NYC

[Tyler Wright]

I had hopes about reflex:

From last MAX or 360|Flex last March? 6 months later ...

 

"Introducing Reflex" - "We hope to provide a compelling alternative to Flex that is smaller, faster, and easier to build on top of."

"What project will replace the Flex framework?"

Also there was a promise to have a component set at least as large as the one from Flex 3,

collaboration with FDT team

discussions with degrafa team to have a release for reflex

RPC services

These are my hopes too! Adobe has teams of full-time developers and project managers who got them where they are. We're a few smart guys who still think we can outdo them, but a 6 month timeline is pretty ambitious.

This isn't the end, it's only the beginning. True - opensource has risks. While you're investing emotionally I'm taking weeks off of paid work to just work on the project. I hope we both win! :)

In fact, I plan on it.

Tyler Wright

[Jacob Wright]

I agree with both. It is frustrating to work with Flex sometimes (or a lot of time?), and I really had high hopes for a lot of projects presented. I also feel bad when I only push hard on a project in order to present on it at a conference and then let it go. Shame on me for getting others excited about it and then not seeing through what I told them was going to be coming.

But of course, it's easy to ask how one might help out and contribute in lots of ways, beyond just coding.

Jac

[Tyler Wright]

By the way, I understand the disappointment expressed - a "breakup" seems to spell doom for the project. Nothing offensive has been said.

I'm just politely disagreeing. :)  And because I might know a little more about what's really going on with the projects, I might not be wrong.

you all rock by the way! It makes me happy that you care so much about the project(s).

Tyler

[Jacob Wright]

This isn't the end, it's only the beginning. True - opensource has risks. While you're investing emotionally I'm taking weeks off of paid work to just work on the project. I hope we both win! :)

In fact, I plan on it.

Me too. Be sure to follow http://groups.google.com/group/flight-stealth to keep up on the progress!

Jac

[Vasile Ermicioi]

I am not frustrated by reflex: 

- its development is not nearly stopped, 

- you guys are awesome, just seeing on what projects you worked make me feel confident about reflex (flare, flight, degrafa, etc)

..just the thing that Ben will go into another open source project a bit scared me ..

I use flex since version 2 and what I don't like about is that  its size is always growing even for an empty application (flex 2 empty app ~ 170kb, and flex 3 ~ 200kb, and flex 4 > 200kb - release version, merged into the code)

Reflex seems to respond perfectly to that need

You have my greatest appreciation for all your work - THANK YOU!!!

[Ben Stucki]

Thank you to everyone that responded.

Thank you also, Vasile, for your response. I know it's hard to work on a moving target, but I hope I'm able to see the skins you're working on soon. I could probably use them for an upcoming release.

Speaking of moving targets, there are going to be a lot of changes going on in the repo over the next few weeks. I'm planning to follow Tyler's changes in Stealth up until he makes his presentation next week, and will then begin development towards an independent Reflex release in early October. For those that are interested, I've posted a Reflex roadmap on the official site. Note that this roadmap is separate from Tyler's Stealth project, but maybe he'll be able to keep up anyways. :)

I hope this shows that I absolutely do intend to continue developing Reflex, and I think that you will see a lot of progress in the few weeks following 360 and up until MAX. Still, this roadmap only represents my current best intentions, and it might change with time. If someone offers me a million dollar contract to build their own internal component set for example, I'll probably disappear for a while. I'd like to think that I can build Reflex to a point where the community would continue its development after I'm gone, but it's going to take all of your help over the next several months (even years) to do that.

- Ben

[arieljake]

1. Can you guys explain the ideological differences causing the split
between reflex and flight?
2. Can you guys spend a little more time developing pages on your
sites describing your plans for your projects, your immediate needs,
and developing some documentation to get us up to speed on the
existing code base? I have time and interest in contributing, but it
is not clear to me where and how to start to support the effort rather
than hijack it, and while your code may seem obvious to you, when I
read it I don't find many small comments explaining your thinking.

[Robert Penner]

+1

[John Lindquist]

@arieljake - I'll summarize #1 to save Ben and Tyler the hassle since Tyler sits behind me and talks my ear off about it

Reflex - ideally architected and ultimately extensible
Stealth - approachable and user-friendly

You should hear the arguments they have over whether a "Position" class should have a percent member for ease-of-use or whether the percent calculations should be broken in a utils class for architecture. :)

I'm sure they'll both share a bulk of code, but there will definitely be implementation differences. I guess the community will decide the winner.

*disclaimer - Ben and Ty may feel differently than my summary

[Ben Stucki]

I do feel differently about John's summary.

As a blanket statement, I would say that there aren't many (if any) ideological differences. We both want a light-weight, easy-to-use, Flex independent component framework, and we agree on most of the large concepts around using behaviors and skins etc. The differences causing me to split were simply that the devil is in the details, and we've come across a large assortment of differences there. What I'd like to do (as opposed to filing a list of technical complaints back and forth on this email thread) is promise to point out any major technical differences between Reflex and Flight as they are being committed and the projects begin to diverge. This way I will be accountable to explain my reasoning and everyone will have a forum to weigh in. Fair enough?

For #2, I'm working on it - and did you see the roadmap? :)

- Ben

[Jacob Wright]

Tyler is heads down right now working hard to get some things done for 360|Flex. He has put together a roadmap for Stealth and is putting together a document with the vision, architecture, and needs for the project so that people may help out.

And I agree with John's summary. The one is more idealistic and the other is more pragmatic (or less idealistic anyway). I think both projects will move forward now that those devilish details can be worked out independently. :)

Jac

[Ben Stucki]

I'd like to reiterate that I intend for Reflex to be easy to use. I also believe Reflex's "idealistic architecture" (I prefer simplistic) will be "satisfactorily fit for use" - which does, as I understand it, make it pragmatic. If a large number of people find Stealth's approaches easier to use AND it doesn't overcomplicate Reflex's architecture I will happily steal them - but simplicity in design (ie. architecture) continues to be my goal. I believe that this is the source of what will make Reflex easy to use.

Also, while these lofty ideas and antagonistic remarks start to fly let's keep in mind that both of our code bases still consist of a non-inconsequential segment of each-other's code, idealism, pragmatism and the like.

- Ben

[Mrinal Wadhwa]

Ben, Tyler, Jac and everyone else who's been contributing ...

I welcome the change, forks are good, I've learned a lot from the reflex codebase over the past months and it was fun observing how you approached various common problems in component/framework development .. a fork only means that we'll see multiple approaches being implemented and that's great ! gives me more to steal from ;) .. grateful that you guys are spending your valuable time sharing such high quality code with an open license and wishing both the projects tons of success.

Cheers,

Mrinal  

[Vasile Ermicioi]

http://blog.flexgeek.in/2009/11/somethinglikemac/  translated to degrafa and used with reflex,

not finished but some of them are usable

release is less than 120kb

also I added FTELabel from an older reflex release :)

[arieljake]

You guys can do what you want. But I think the fork is a mistake. You
are not only cutting your organization in half, you will ultimately be
cutting your open source communities that will support development in
half, over some small differences. From the sound of it, this is a
political issue not a technical one, and if you want your projects to
be successful sometimes you have to deliver a sub-ideal product for
the sake of reality. We are all talking about an entire framework
being developed, who cares if you have a percentage setter or not. If
it means keeping one of the twins on the project so be it. You can't
tell me that any of you expect to
build a flex replacement without some frustrating compromises. In the
meantime, while the main developers will be happy that they have
retained control over their respective projects in a very geeky way,
the community will be stuck choosing between two incomplete efforts.

>  degrafa_mac_skin.zip
> 1049KViewDownload

[John Lindquist]

I think the competition between frameworks will drive them to success.

And they'll both "borrow" from each other's progress without getting hung up on details.

Sounds like a "win win" to me.

On Sep 18, 2010 10:54 AM, "arieljake" <arie...@gmail.com> wrote:

You guys can do what you want. But I think the fork is a mistake. You
are not only cutting your organization in half, you will ultimately be
cutting your open source communities that will support development in
half, over some small differences. From the sound of it, this is a
political issue not a technical one, and if you want your projects to
be successful sometimes you have to deliver a sub-ideal product for
the sake of reality. We are all talking about an entire framework
being developed, who cares if you have a percentage setter or not. If
it means keeping one of the twins on the project so be it. You can't
tell me that any of you expect to
build a flex replacement without some frustrating compromises. In the
meantime, while the main developers will be happy that they have
retained control over their respective projects in a very geeky way,
the community will be stuck choosing between two incomplete efforts.

On Sep 18, 8:22 am, Vasile Ermicioi <elff...@gmail.com> wrote:

> http://blog.flexgeek.in/2009/11/...

>  degrafa_mac_skin.zip
> 1049KViewDownload

[Jacob Wright]

arieljake does have a good point. That was my concern. If someone is wanting to contribute, which framework do they put their efforts behind? But I think the reason for the split was because the many discussions about things like percentage setters (on all sides, not just Ben's) have been holding the project back since the beginning. Compromises have been made throughout, but it's taken quite awhile to get where we are. I believe Ben and Tyler are just trying to do what's best for the project.

And sorry for any antagonistic feeling remarks Ben. There was a question asked about why the split and what the differences would be in the two projects. That's hard to answer without bias and with complete tact. I apologize.

Jacob

[Ben Stucki]

Dear arieljake,

There is a sub-ideal, compromising, pragmatic component framework available for your use. It's called Flex. I think it's quite good, and I use it almost every work day. In fact I've learned quite a lot from it, but I'm not trying to reproduce it. In addition to the good bits, I've learned from all of the frustrating pain points that I (and others) have felt when creating real world applications with it. I'm endeavoring to make something better, something less compromising, something less frustrating, something more ideal.

I do expect to build a Flex replacement without frustrating compromises.


- Ben Stucki

P.S. Saying it was one thing you thought I couldn't do. Now I'm keeping count.

[Ben Stucki]

Thanks for the support Mrinal! I think your comments are spot on (and very much appreciated).

Also, thanks for the follow-up John and Jacob. I agree it's a win-win, and I'm pretty excited about having an all-clear to forge ahead. I'm sure Tyler is too.

Finally, thanks for sharing Vasile! I'm not sure if I'm ready to use Degrafa for the default skins, but this will definitely help me make sure I build the right pieces into Reflex's graphics primitives. Hopefully I can make creating skins like this a little easier for you in the next few releases.

- Ben

[Vasile Ermicioi]

I developed as fast as I could and here is a first demo of my version of reflex graphics (no more degrafa)

credits will add later now is just a demo

it includes primitives: Rect, Ellipse, Line, Path (Path code stoled :) )

and a few demos of usage includes a few Aqua skins and spark skins

and that for 140kb in debug mode

release mode seems to not work, I guess compiler deletes metadata which reflex is using

later this week with more complete skins

[Ben Stucki]

Hi Vasile,

I just wanted to say thanks for the updates. I'll be taking a closer look at them soon and will see where I can add the contributions in (with credit where credit's due of course). The primitives sound like they belong in the framework repo. I might also start a reflex-lib repo soon as an area for the community to share custom components, behaviors and skins (like these). Again, thanks for sharing! I think it's helping to point me in the right direction.

- Ben

[Vasile Ermicioi]

hi again,

I refactored a bit:

- primitives are based on Shape not on Sprite (this way I can use invalidation)

- only 'draw' method should be overrided

- 'forked' SlideBehavior and StepBehavior - changed to have bindable properties

 It is amazingly easy to develop components, thank you guys

Now I can tell you a few things about reflex:

Features that I love:

1) Component architecture thinking about it as Skin+Behaviors

1) Binding and Bind tag to use binding from mxml

2) ButtonEvent and behavior

3) Invalidation (RenderPhase.invalidate)

4) Layouts

What I am missing:

1) 'disabled' state  - Tyler already added it on flight-stealth ('enable' property)

2) Flex like states to use in skins (and may be transitions)

3) Position classes and SlideBehavior and StepBehavior seems a bit broken (at least not bindable)

The biggest problem that I am facing is that code is not working if exporting in release mode

[Jacob Wright]

The biggest problem that I am facing is that code is not working if exporting in release mode

I believe that is because in debug mode ALL metadata is kept. Release mode only metadata included in the compiler settings is kept (-keep-as3-metadata += Bind,BindProperty,etc as far as I can remember, somebody help me out). These aren't needed in your final project if a SWC you're using had them in their compiler settings.

Hope that's it. It's an easy fix.

Jacob

[Vasile Ermicioi]

Hope that's it. It's an easy fix.

Yes, thats it, thank you

-keep-as3-metadata+=Style,Bind,BindProperty,Binding,CommitProperties,EventListener,LayoutProperty,PropertyListener

and  

here is a reflex demo of 74 kb size :)

with 2 button skins, vertical scrollbar, vertical and horizontal slider, a checkbox and 2 radiobuttons 

and primitives : Rect, Ellipse, Line, and Path

+ reflex builtin Group, VGroup, HGroup and a few behaviors

I  u reflex !!!

[Tyler Wright]

Way to own it! You're totally rockin the Reflex Vasile.

Tyler

[Baz]

Vasile that's great! Source? :)

[Vasile Ermicioi]

for source look at previous posts, 

just add the compiler option to keep metadata, needed only for release,

if reflex lib is not with sources take it from http://github.com/benstucki/reflex