Huge Adobe Partnership to Open Source Flex with Apache Software Foundation …

… Could have been the title of this post 3 months ago. Given the recent layoffs and announcements in the past few weeks, it may seem short sighted to be overly excited faced with this news today. This was, however, a part of the announcements recently by Adobe regarding Flex. They are committed to contributing Flex (and several complimentary projects) to the Apache Software Foundation (ASF). I personally witnessed the commitment and plan for this at their San Francisco office. Many leaders of the Flex community gathered at Adobe SF by request of group product managers Deepa Subramaniam and Andrew Shorten this week. We heard from senior management on their new strategic vision for Adobe around digital marketing and digital media. We heard apologies for the accelerated and miscommunicated nature of the messaging of the past few weeks. And, we, most importantly, heard about the future of Flex as it may likely live in the ASF.

There has been a lot of focus lately on the poor messaging to the community around the announcements of late. I don’t disagree about how destructive and misguided this messaging was, but I do think it’s important to see it as the past and move on. Many in the community may criticize this idea. It would be easy to say this is one of many moves towards under valuing the community. Again, I wouldn’t totally disagree, but I do think this particular moment in history is different than in the past. Here’s a bit of an explanation…

The ASF model for development is much different than how a product is developed by a corporate software vendor. On an Apache project there are no project managers, product managers, senior vice presidents, presidents, CEOs, CTOs, development managers, senior engineers, or any other title of a role that you can think of; other than “commitor”. A commitor has the right to commit code to and vote on a project. You can read a lot more about the ASF process on their site. So, what’s the big deal? Why is this different than a time in the past that Adobe (or any company) did wrong by it’s community, said “we’re sorry”, and kept moving right along? – This time they’re actually handing over the reigns of the project to us; to the community.

Adobe will not have contributors on the Apache Flex project. Neither will Spoon, Roundarch or any other corporation or group. This is (technically) true because contributors are individuals. So, there will be some individuals who are contributors on Apache Flex that will also be Adobe employees, but they will not have any higher status or privilege than any other contributors on the project. This is by design. The design of the ASF. The process of Adobe contributing Flex to Apache is in the form of a proposal. That proposal states the intention of all initial commitors to create an Apache Flex project, among other details. Up until Apache Flex is realized as a full project there will be a phase of incubation. During this time Apache will work with the proposal submitters to help arrive at a proposal that fits the Apache model. This is to ensure the best possible arrangement for a successful Apache Flex. That’s why this moment is different than similar times in the past. By this model, Adobe gives up the sole control and direction of Flex and enters into a partnership with the community to drive the future of the framework, with the guidance of Apache. Without that partnership Apache will not accept this contribution. So, for those claiming Adobe is “giving Flex to Apache to die” or “dumping off Flex with Apache”, this is simply not true. It also shows a severe lack of understanding and respect for what the ASF does and has done.

We have not heard this before. This is a new approach with a high potential for excellent collaboration of all the individuals and entities that depend on Flex for critical business needs. These individuals and enterprises need a path to a future that enables their teams and businesses to deliver the high quality solutions they (and their customers) are accustomed to. For many enterprises developing highly complex applications with millions of lines of code, large development teams, and high expectations for success, Flex was and still is an extremely valid technology, for others it is a requirement. There are enterprises that are so vested in using Flex to build (or maintain) a platform for their businesses that a short term migration is simply not possible. This is not an understatement. Individuals from enterprises like what I’m describing were present and vocal at this recent Flex Summit.

So, what else did we hear as it’s relevant to Flex?

A paraphrased quotation, as I can’t remember the exact wording:

Adobe will not continue to develop Flex as a “stand alone” business.

It was also stated that Adobe does not currently have plans to develop any product targeted at enterprise scale user interface development. That’s not to say they aren’t developing tools for HTML5 development, just not a platform like Flex based on HTML5. They are committed to their runtimes for what seems like a minimum of 3-5 years. All of the current runtimes (less Flash Player for mobile browsers): Flash Player for desktop browsers, AIR for mobile (iOS, Android), and AIR for desktop (Windows, Mac OS).  However, it does seem clear that this runtime development will be focused in line with their new strategy of digital marketing and digital media. For Flash that means game development and video. There are certainly concerns around this regarding Flex. Game development is very different from enterprise class development. They have also stated that they will commit to backwards compatibility to meet all the dependencies that exist today in Flex 4.6.

My thoughts on the future of Apache Flex…

Apache is the the future for enterprise class Flex. For those highly skeptical and critical of Adobe, my message to you is this: The move to Apache is a big one and categorically different from anything we’ve seen in the past. Is Apache perfect? I doubt it. What it’s not is proprietary or corporate. It also answers in no way to a group of public share holders. It is not driven by senior management who are making business decisions to please those public share holders. It has no “layoffs”. It is not driven by the sales of an IDE. Further more, it has no goals of making profit – at all.  It was founded and is governed by folks that really understand open source development. It is one of the most respected software foundations in the world. What Apache Flex can be is a powerful force for the community to drive the future development of Flex and deliver in a way that holds true with the most important needs of the community – governed and guided by the Apache model for open source development.

I also believe that development is just one factor in the future success of Flex. Now that Adobe is stepping back from Flex there will be gaps in the future in the areas of: education, community (user groups, conferences, etc), quality assurance, support, tooling and potentially others. The Spoon Project may likely provide an appropriate group to support in these areas. There are already of mass of talented folks rallied around problems similar to these gaps in Spoon.

So, down to brass tacks:

What’s the reality of HTML5 as it relates to Flex at this moment in time?

HTML5 and open standards for the web will play a dominant role in the future of technology. What I believe is that this future is farther off and a bit more uncertain than many that make this statement. Let’s consider some recent figures by IDC. IDC predicts that 90% of smartphones and tablets will have HTML5-ready browsers by 2013. This means that 90% of consumers will be able to access your HTML5 application on their smartphone or tablet by end of 2013 (still 12 months away). However, the Gartner prediction of 81% smartphone penetration/adoption for 2012 was off by 50%. That stat is actually 44%. Even looking at the past month, smart phone purchases are only at 56%, well below that 81% prediction. The staggering number is 2015. 90% browser adoption of HTML5 is estimated for 2015. The Flash Player has historically had 90-99% adoption in desktop browsers. Additionally, there is no complete HTML5 specification. Notice: “Draft”.

In that respect an even more staggering number is 0. 0 is, technically, the current browsers that support an open standards based implementation of HTML5. Without a complete open standard it’s impossible to have compliance around that standard.

As Google, Microsoft, Apple, now Adobe and others seem to have consensus around the support for HTML5 in the future, none of these companies seem to do a great job at messaging when, exactly, this future is. Along with that I think many analysts are also doing a poor job of analyzing and messaging the capabilities of current web technologies. Part of that problem is the term HTML5 is now being used for anything related to HTML. Current support for HTML, JavaScript and CSS is a lot better than it was years ago, but comparatively (both to what HTML5 will be and what Flex and the like are now) it has many drawbacks. There is still a considerable amount of browser fragmentation to deal with, there are considerable challenges with styling for some experiences, there is a severe limit to the amount of performance of any browser’s JavaScript “runtime”, and there is still a lacking of support for mature approaches to software engineering when it comes to JavaScript. That doesn’t mean I hate JavaScript or think it isn’t a great tool for solving a lot of problems. It is my opinion on the comparison of the available choices for developing software on the web today – not the how that may or may not look in the future.

In the future that is next week and next year, Flex still has a place.. and that future is considerably more relevant than speculation on what is 3 years away… but back to the topic at hand: Apache doesn’t hedge on those futures. It supports a thriving community interested in building software. This is another huge reason that I see so much potential for Apache Flex. So, you won’t be hearing me making blanket statements like HTML5 is the future and you shouldn’t be building Flex apps. Or, Flex will be around for decades and HTML5 is completely overrated. Anyone who makes makes these kinds of huge generalizations is in one (or all) of these categories:

  • A person who is trying to sell you something
  • A company man (or woman)
  • A technology zealot

If you’re an engineer or part of management driving the technological future of your organization, it’s probably not a good idea to listen to these kinds of people as your only source of information (or maybe at all). The sales person is driven by profits. The company man is driven by a career towing the company line. The technology zealot holds to their particular technology views without necessarily analyzing the reality of the tech landscape. These are all risky people to put your faith in. The technologist that can analyze trade offs, discuss options, strategize to meet business goals effectively; that is a person who I trust. That is a person who I want on my team, working with my customers, and solving critical technology problems with me.

More to come on Apache Flex as it develops…

Google+ comments.

40 thoughts on “Huge Adobe Partnership to Open Source Flex with Apache Software Foundation …

  1. Masu

    Flex life would be easier if @flashcatalyst and the “Design View” in @flashbuilder would have survived! This is a big “hit in the face” for everyone planning to build large-scale Flex apps in future :(

    How should we create rapid #Flex prototypes for our customers now? How should we glue Photoshop design skins on Spark components? Are there any alternatives to @flashcatalyst? How much is the @flashcatalyst source code?

    1. adamflater Post author

      Interesting, Masu. I’ve never really used Design View or Catalyst much for production enterprise work. For me, coding MXML has always been faster than a WYSIWYG tool, but I guess that’s not true for everyone.

        1. adamflater Post author

          Sure.. people learn differently.. my opinion about Flex is that it’s a tool for programmers. In my experience WYSIWYG tools lack a level of quality that you can depend on for enterprise level code management.

      1. Masu

        Ah yeah … and this will definitely effect Flex beginners learning Flex … Removing those visual features (from Flash Builder) and tools (like Flash Catalyst) will add one more barrier to Flex learners … and this is bad because we need more highly skilled Flex developers around.

        1. adamflater Post author

          I agree with that.. again, my opinion: but things like design view actually create bad habits and not highly skilled developers.

        1. Brian Kotek

          The reality is that sales of Catalyst didn’t come anywhere close to offsetting the cost of creating and maintaining it. So unfortunately it’s extremely unlikely that they would bring it back to life.

    2. Praveen

      I too in my Flex devlopment career didn’t used ‘Design View’ much. For 90% there won’t be much difference if there’s or not but for beginners it’s very useful

  2. David H.

    This an excellent article. I think that Flex being owned by Adobe kept a lot of companies from using it in a larger capacity, though I’m sure many usedit internally. Maybe they just weren’t willing to pay adobe for it. Hopefully now that it is open sourced it will be more viable for companiesto use, particularly if html5 doesn’t pan out so well. I don’t really like this becauseI always felt that Adobe deseved to be paid for this software. Anyway, we’ll see what happens.

    1. Gazinhio

      I agree that making Flex open source should remove a lot of barriers to entry. The trouble I see is that the development SDK is open source but the runtimes are not. Therefore the Apache Flex community can do all it wants but if Adobe move Flash and AIR in a different direction then there’s trouble. Their announced intended support for AIR for Metro apps on Win8 does give me some confidence. How about on Windows phone?

  3. Gustavo Aquino

    I don’t believe on your words.

    Lets wait 6 -12 months to come and see what’s happening.

    My guess is Adobe is working on next killer html5 sdk to compete with JQuery.

    1. adamflater Post author

      Gustavo – Well, that would be speculation. I’m just reporting what senior management relayed to us.

    2. Steve

      Nope, At the Flex Summit Adobe announced they are not creating an html5 application equivalent to Flex. Did I hear that wrong??? Edge is the closest thing but that is for creatives. We are on our own if we want to create enterprise level html5 apps. I doubt they would compete with JQuery, they are currently integrating JQuery into stuff…

  4. Shashank Tiwari

    Apache is great at nurturing young projects. Time will tell if they can revive dying one’s too. Adobe made sure through multiple mess-ups over the last couple of years that a smoothly flying Flex airship crash landed in the middle. The growing HTML5 wind made things worse. Now they seem to have hit SOS button and it seems Apache will help provide the parachute. Good luck. I would like to see Flex survive (though, realistically the chances appear slim given its downward wobbly decent is underway!) and Adobe learn to focus on its strategic directions (which seems to change as often as Google’s Doodles these days).

    1. adamflater Post author

      Shashank – I definitely respect your opinion, but I think assuming Apache would adopt something in the way you’re speaking is a bit naive. Naive on the level of quality and guidance they give to projects. I don’t think anyone is suggesting is a shoe in to bring Flex back to the trajectory it was on 4 year ago. That doesn’t mean there is no place for a technology like Flex and it does make it less of a battle of Adove vs. X now. Now the goal is how well the community can support and develop the product for itself.

        1. Amit Goel

          Thanks a lot. Nice to read this when Adobe has created a havoc over Flex. Going to ASF, Flex community is only going to grow more here after.

          Those are all crappy comments put their by some of the Flex hater kind of nerds. Doesn’t count at all. Don’t expect any deja vu here.
          Flex is purely a data presentation framework, with large community support. It’s not an animation creating tool which instead Flash was. Being a de facto RIA standard it will remain a great tool and cool framework, and it will stand in the market until HTML5 comes up with a solid specs and more specifically on browser compatibility. I hate(obviously everyone) putting those IF..ELSE checking browser vendor & even version(OMG) all the time while writing my pages.
          Sure Incidences are Functions of Time – we’ll have to wait and see.

        2. adamflater Post author

          Interesting opinions… btw, that content was probably served by Apache Web Server.. and I’m guessing you’ve used ANT this week at least once.. yeah?

          1. Shashank Tiwari

            I am not here to either argue or to convince you. You know me well and hopefully will agree that a man who wrote 3 books on Flex, ran some of the largest Flex Camps over the years, had a ton of Flex clients, and possibly was involved in some of the largest Flex projects, isn’t a Flex hater! In fact as you may possibly guess, I quite like the framework. The problem though is that my liking isn’t enough or relevant to Flex’s future. The enterprise world always had difficultly including Flash related tools in their development stack. Flex broke that barrier and was well adopted for a bit. Then our dear friends at Adobe screwed things up first by trying to rebrand it, then trying to get entangled into the HTML5 debate, and finally sucking up to the promise of HTML5. Now Adobe has, through their actions, proclaimed Flex to be terminally ill and incurable. The only fate of such a technology usually is marginalization.

            Now as far as Apache foundation goes, they are one big warehouse of good and bad projects. The Gold standard that Apache was has been diluted for some time now. Some projects are very successful, some are moderately used, and a few are simply junk. Most successful Apache projects in the recent times, for example Hadoop family of tools, or Cassandra, have seen huge infusion of capital and tons of full-time paid developers working on projects. Much of this money is being pumped in because the investors have a plan for monetizing and earning multiple times of their investment. Apache projects or for that matter most successful open source projects are no more run by hippies who code all night for love and fresh air. In contrast to the well funded Apache projects, the already limited Flex community is thinning out everyday and most of the folks are members of small consulting firms. Consulting firms usually follow the direction of the wind, unless they want to go out of business. Given that how many firms are going to put their money, reputation, and future in the medium to long term in trying to save Flex. Possibly a handful will. But most of these companies, despite the noise they make today, are so small (where there revenues are often less than 15 to 20 million dollars a year and they are not sitting on million of dollars of VC money either) that I wonder if they can afford to swim against the tide.

            Last but not the least, have you wondered why Roy Fielding is singing songs of Adobe praise and trying to sell Apache to the Flex community and make it appear like its for the sake of freedom and prosperity that Adobe is donating Flex to Apache. (Flex is already open source. Isn’t it? BlazeDS is too, which apparently hasn’t seen updates for sometime now :) ) With all due respect to his talent and his contributions to open source, its quite possible that the $250 million that Adobe paid to acquire Day is what is making him speak.

            Good luck. This is the last comment from my side on this thread. Pardon me if I hurt your sentiments. My intentions wasn’t to do so and I will only be happy if Flex gets some life support cause without it Adobe has left it to die. Also, if any current Flex developer desires to remain decently employed in the longer term accepting the reality and looking beyond Flex could possibly be helpful. Alternatives are behind today. They may not be tomorrow. Also, I would urge that try not too be fooled again by the HTML5 drama Adobe is trying to sell or you will be helping incubate the Edge project a few years from now as well ;)

  5. Pingback: The Future of the Flex Framework in Enterprise IT « Yakov Fain's Blog

  6. JCLang

    Nice post.
    But I still don’t understand the idea of “HTMLx application”.
    I mean: what does that mean ? If I use/want to create/think of an application, I certainly not think of an “evolved” web site. Specially on Smartphone/tablet where perf. are at stake.
    What I need is an optimized runtime, with a cross-platform capability (including mobile), and an oo language based on a solid backbone, with some (optional but so comfortable) wysiwyg UI builder: well the good news is that I *have* it today. Will I move (back…) to HTMLx/JavaScript stuff? well, for now I focus on real apps (not ironic).

    1. adamflater Post author

      There are JavaScript frameworks developing built on HTML5 (brunch, for example). They will surface to be more mature around building the kinds of “apps” you’re talking about. However, there’s still a fundamental problem (in my mind) that this technology is built on mark up. Yes, Canvas and SVG, etc change some of this, but mark up is what these technologies are built on and mark up is not a particularly nice way to engineer user interfaces.

  7. Pingback: Huge Adobe Partnership to Open Source Flex with Apache Software Foundation … | Adam Flater : Technology Blog | Everything about Flash |

  8. Raju Bitter

    Interesting article. Although the statement that 0.0 % of all browsers support HTML5 is misleading. From a technical perspective, there enough good workarounds to make up for the differences in browser support. OpenLaszlo, another Flash/JavaScript based technology (originally only SWF), created a “DHTML”/JavaScript runtime back in 2006/2007. In 2009 I saw their massive Laszlo Webtop application (110k+ lines of XML/JavaScript) code run as both an SWF and JavaScript powered application, rendering almost pixel-perfect in both environments.
    Here’s a short video showing Webtop, just to give you an idea

    OpenLaszlo compiles XML and JavaScript to JavaScript 1.5, or ActionScript 3 (for SWF using the embedded Flex compiler to generate the SWF file).
    That’s exactly the kind of technology we need to make a smooth shift over to an HTML5 “only” web, over a period of 2-3 years for the web (and longer for enterprise/legacy applications). The only problem
    Flex has is that it the components are relatively heavy-weight. As a result, the first FalconJS demos are reported to be 5.5 mb in size for a simple demo app (uncompressed JavaScript code), for OpenLaszlo a similar app is less than 600k of JavaScript code. If the Apache community (including Adobe) would manage to come with a light-weight version of Flex, that would be a major improvement. In general, I think that SDKs are better managed through a community process than by the product managers at one single company. It’s going to be interesting to see what the community wants to turn Flex into…

    1. adamflater Post author

      That 0% statement was in relation to what people are saying about an open standards based web and that the particular standard for HTML5 is still not completed.

      Also, remember: The current state of Falcon JS is R&D. This was a research project to learn about the feasibility of a concept like what’s currently implemented in Falcon JS. No one is suggesting that this is a production ready tool. It’s also only one path of several possible solutions to supporting cross compilation to HTML/JS.

  9. Pingback: GradleFx automated builds | RIAgora

  10. Pingback: Flex and It’s Future as an Apache Project | Roundarch Blog

  11. Edwin

    As one of many who spent a non-trivial amount of time and money hearing Adobe inform and make their case for the Flex platform (enterprise and mobile) in Oct 2011 only to see them blow up that platform a few weeks later, it’s still hard to blow off the WTF of it all.

    Adobe is killing it’s Flash platform so that it can get into digital ad metrics? While it’s still hard to experience the web circa 2011/2012 without encountering lot’s of Flash? That just seems crazy.

    If Adobe had delivered the whole Flash platform to Apache (Flash/AIR/Flex) instead of just Flex and done so in a way that respected the commitments many of us have made to the platform, this might be different. But it’s hard to see how Apache Flex has a future when Adobe will probably blow up Flash/AIR in another clusterf__k sooner rather than later.

  12. Tom

    Remember Aldus? Remember GoLive? It was a really cool and different HTML editor until Adobe trashed it after (Macromedia) DreamWeaver came along and took all the marketshare. Macromedia understood the importance of its user community and building an ecosystem until Adobe swallowed it. Then there was Scintilla CoolEdit, an awesome audio editor which Adobe pounced on and wanted $300.00 U.S. for putting their logo on it. I think this is the best possible outcome for Flex. Viva Apache!

  13. Graham Phillips

    All well and good, but the Flash Player, upon which the Flex SDK runs and depends, is still and forever will be (as long as Flash exists as a product) an internal Adobe product. No-one has yet explained how the Flex SDK and Flash Builder (which although Adobe have pledged to continue to support, I believe they will be dropping within the next 2 years) will remain in sync with developments and innovations in the Flash plug-in. They also want to eventually open-source the compiler, thereby washing their hands of the whole technology. However you spin it, Adobe have taken a strategic decision to ditch “enterprise” and focus on marketing which has left a lot of us in the lurch. HTML5 and iPhone have been used as the excuse.

  14. Pingback: Adobe Flex se mueve a Apache « Nono F. Carballo Escalona

Comments are closed.