Comments for Flash Builder, Linux and Winning the RIA War (http://www.insideria.com/2009/06/flex-builder-linux-and-open-so.html)

Flash Builder, Linux and Winning the RIA War

2009-06-18T14:00:00Z

Summary: Adobe should address developers with a marketing program that is separate from their designer marketing program, and offer products aimed at hard-core developer interests. The goal would be to leverage the influential developer community to drive sales of products that designers and other non-programmers purchase, not to maximize ROI for developers. At stake is RIA domination and sustainable and profitable business for Adobe, its partners, and its customers. I submit that the converse - failure to engage effectively with developers and to recognize the strategic value of the developer market as a loss leader - would result in losing the RIA war and making the rest of Adobe's products vulnerable in a way that they are not at present.

I attended FlexCamp at Adobe's San Francisco premises two weeks ago, where I spoke with numerous Adobe employees who work on the product formerly known as Flex Builder (now rebranded "Flash Builder"), Flash Catalyst and Flash SDK, including engineers, managers, product managers, a director and two VPs. The following is a restatement of what I told Adobe staff.

The Adobe Flash Platform provides a first-class result: web-enabled applications that present well and are a joy to use. Unfortunately, because of the weak development toolchain, the cost and time to develop applications based on Flash are much higher than what they should be. Microsoft's excellent toolchain provides a significant advantage for Silverlight. Developers working on Silverlight can be many times more productive than similarly talented developers working with Flex. Customers are learning that projects built with the Adobe Flash Platform take significantly longer and cost more than similar projects built with competitive RIA technology.

I am an independent software developer, and my hourly rate is relatively high. I justify this by focusing on the effectiveness by which I generate value for clients. In addition to gathering requirements and customer feedback, I spend much of my time writing, compiling and testing Flex and Java code. I must continuously invest in software tools and more powerful computing platforms in order to remain competitive, as well as employ appropriate methodologies and technologies. I maintain my own servers in a data center, so customers can play with applications as they are developed in a live setting. I prefer to develop on Linux and deploy clients to Windows, Mac and Linux, and deploy the back-end to a Linux server.

Given the appropriate hardware platform and a suitable application, a 64-bit machine with 64-bit software can run considerably faster than the 32-bit equivalent. For example, the Java back-end for a project that I am currently working on takes over 2 hours to compile on my Dell D620 laptop (Intel Core 2 Duo, 2GB RAM, Windows XP), 20 minutes on the desktop computer I built three years ago (Intel Core 2 Duo @ 2.1 GHz, 4GB RAM, Windows XP) and less than five seconds on the 64-bit computer I built last month. This Chevrolet-priced computer performs like a Porsche. It is amazing what $600 will buy today if you don't pay the Microsoft tax: a 3GHz Intel E8400 Core 2 Duo processor, a 1TB drive, and 8GB of 1333MHz DDR3 RAM on a SuperMicro C2SEA motherboard with integrated video, running the 64 bit version of Ubuntu 9.04. 64-bit Windows has issues that limits its usefulness, and 64-bit Mac OS/X benchmarks show performance within 10% of 32-bit OS/X. Published benchmarks between 64-bit OS/X and Linux are interesting, but I am only interested in how fast I can build Flex clients that interact with Java servers. The only benchmark that matters to me is how fast Eclipse works.

Eclipse is a terrific IDE for many purposes, including server-side Java development. Unfortunately, Flex/Flash Builder, whether installed as a standalone package or as an Eclipse plug-in does not demonstrate similar maturity. One might expect that a name like "Flex Builder 3" or "Flash Builder 4" implies a mature product, but that is misleading. Given a Java project of reasonable complexity, and a Flex project of similar or less complexity, you will soon recognize that compile times are significantly longer for Flex/Flash Builder, and many features on which Java programmers rely on daily are not implemented for Flash Builder, are implemented poorly, or have significant issues.

Flex Builder 3 was released in the summer of 2007 when Eclipse v3.3 (Europa) was current. A few minor releases were issued since then, but without any new productivity features. An alpha version of Flex Builder for Linux was released last August, but no further release dates have been announced. Flash Builder 4 has only had 32-bit versions announced for Windows and Mac. Flash Builder 4 for Windows just went to beta and release is expected in Q4/09. A product cycle over two years in length means that the product is out of date with the other software that it must interact with with for most of its life.

It is no secret that the Flex compiler is very slow. This impacts the edit/compile/debug cycle that developers repeat when writing and debugging code. On a large program, it may take 5 minutes to make a change, compile and deploy the software, then run the program to the state that needs to be debugged. Developers loop through this cycle dozens of times each day. When Eclipse is used with Java, a feature called hot code replace allows the Java program to be paused at a breakpoint and a small change made to the program before resuming execution from the beginning of the stack frame. Hot code replace short-circuits the edit/compile/debug cycle, thereby providing a significant productivity boost. Adding hot code replace to Flex/Flash Builder would possibly be the most significant single productivity feature that could be added. Of course, a compiler that ran an order of magnitude faster would also provide a big productivity boost. The compiler used in Flash Builder 4 will apparently be somewhat faster, but will not compare to the performance of today's Java compilers or any of Microsoft's compilers.

One key differentiation between Flex and other RIAs has been the cross-platform story. Similar to Java's "write once, run everywhere" message, Adobe has positioned the Flash runtime for Flex as the RIA platform of choice for applications that need to address the widest possible audience. This message does not hold up under scrutiny. For example, the Flash player was only prereleased for Linux as a 64-bit runtime, and the latest refresh was Feb. 10, 2009. The Linux cadre of developers is large, and encompasses many experienced Java developers. They tend to be senior, and are quite influential in the developer community. The lack of solid Linux support from Adobe has been a key reason that most of those key technologists have not adopted the technology. Flex and Flash simply don't work effectively in their primary development environment.

Adobe was originally built on the strength of OEM sales (remember Postscript printers?) and more in recent years has targeted graphic designers with products like Dreamweaver and Illustrator. The PDF franchise continues to do well, and is primarily targeted at business users. Before Macromedia merged with Adobe, their product line resembled the Adobe product line in many ways, and some products competed head-to-head. Although Macromedia purchased JRun, the first commercial servlet engine, it was embedded into Cold Fusion and new feature development has since ceased. Cold Fusion marketing has walked the line between a promoting a tool for quick and easy web development for non-technical people and providing powerful features in a proprietary package. Traditionally, however, neither Macromedia nor Adobe has not addressed the developer market in a significant way. Until recently, this has been a wise decision.

Designers and developers are very different. They differ in their training, world views, interests and purchasing patterns. Designers are right-brained, are usually only semi-technical and purchase software products and upgrades regularly. Developers are left-brained and very technical; it is difficult to sell tools to them. Like most engineers, developers often prefer to build tools themselves, or to use readily-available free tools. Developers have embraced open source because it gives them control over the tools they use. Designers and business users do not generally share the same opinion of open source.

Developing software tools requires a large investment, but is unlikely to generate a direct return on investment that makes business sense. Most successful software tools in today's market support of other lines of business.

Microsoft has made a large investment in their software tools, and has done a good job of integrating their tooling with their other products. If one needs to develop software for a Microsoft platform, Microsoft software tools are usually on the top of the short list of possible choices. Compared to Windows and Office, Microsoft's software tools do not make the company much net revenue, but they drive the sale of other profitable Microsoft products such as the operating systems through system integrators and independent software vendors.
Eclipse is also the product of a large investment of time and money, originated by IBM. Since its inception, Eclipse has had the benefit of many hundreds of millions of development dollars poured into it. Eclipse and a large percentage of Eclipse add-ons are free. The result is a capable, stable and up-to-date development platform that developers have whole-heartedly embraced. IBM's marketing strategy is nuanced, but IBM has influenced the developer community significantly with Eclipse and has leveraged their product and service offerings by building on Eclipse.
Other once-great software tools companies like Borland and Symantec have dwindled, died or terminated their software tools product lines because they failed to leverage the sales of more profitable products by linking developer tools.

Flash and Flex are clearly key components in Adobe's product strategy, but Flex/Flash Builder revenue cannot justify the development expense for the matrix of platforms and editions necessary to address the small but demanding developer market. Contrast this with sales of millions of units of Dreamweaver, Illustrator and Photoshop. Similarly, the Linux market is small compared to the Windows market; however, servicing developers is a strategic move that significantly impacts sales to non-developers in the larger Windows and Mac markets. The Flash platform could be a vehicle to drive sales of enterprise solutions like LiveCycle ES, LCDS, and other yet to be developed products like enterprise portals. These enterprise solutions will require hard-core developers, who expect Linux support.

Given all of this, competitive RIAs are attractive for developers and their clients. It is no wonder that Silverlight 3 is gaining ground:

Silverlight is equally cross-platform as is Flex; JavaFX even more so because it runs on the Java virtual machine; Ajax provides the best cross-platform support.
Silverlight has a much better toolchain, therefore development cost and time to market for products based on Silverlight is significantly less.
Skilled Microsoft developers can pick up Silverlight more easily than they could learn Flex. Even though JavaFX Script is completely different from Java, it runs on the JVM and integrates seamlessly with existing Java code.
As long as Flash and Flex rely on Windows as the primary development platform, Microsoft wields significant influence on the Flash platform and on developers who work with it. During my time at Borland as product marketing manager for Delphi, C, C++ and .NET IDEs, I gained an appreciation for how vulnerable Borland was to disruption by Microsoft. Adobe's current RIA strategy is similarly vulnerable. Diversifying platform support for Flash and Flex, giving Linux equal emphasis with Windows and Mac, changes the game in a way that can't be measured by next quarter's profits.
GWT and other Ajax-based products have strong support from open source and proprietary tools, and HTML 5 is shaping up to be a strong contender over proprietary RIA platforms.

I have the following suggestions for Adobe to win the battle of the RIAs:

Increase the existing engineering and marketing resources for Flash Builder. Again, anticipated Flash Builder sales alone cannot provide direct ROI justification, but sales across the entire spectrum of Adobe's products should more than compensate for supporting this loss leader. This strategy would be akin to meeting Microsoft on their own ground, fighting a war that they are well positioned to win, but more resources in this regard are clearly needed regardless. For example, the May 15 edition of SD Times contained dozens of pages of ads by Microsoft and their business partners, many targeted at Silverlight. There were no ads from Adobe; in fact, I have never seen an Adobe ad targeting developers. Equal editorial space was given to Java, Eclipse and Microsoft, but the term "Flash Platform" was never mentioned. Software won't sell itself, and the competition is strong, well entrenched, and at the top of its game. The history of software is littered with defunct products and companies that had superior technology but inferior marketing.
Initiate a developer program. An empowered cadre of developers steers larger markets. Developers make tools, templates, and components, and end users buy them. Microsoft and Autodesk built their success by fostering a developer ecosystem built around their products. Many successful developer programs provide free product to developers for that reason; many developer products are also free to qualified individuals and companies.
Open source Flash Builder, and build alliances with partners so development can be coordinated effectively. Again, this is not offered as a money-saving alternative - on the contrary, additional resources may be required to support the open source initiative effectively. Note that Zend and their Eclipse PDT (PHP Development Tools) has been open sourced for years, but only Zend engineers do the development.
Outsource some Flash Builder development to a company that specializes in Eclipse development, like Xored.

This is a turning point in the battle of the RIA. Autodesk and Microsoft provide strong historical examples of how to leverage developers effectively, but their stories predate the open source movement. Updating the Autodesk playbook, written in the mid-eighties, to encompass open source will provide Adobe the leverage to build the momentum it needs to win the RIA war, and provide the steady growth that shareholders demand.

Adobe Responds

An Adobe representative who reviewed this article prior to publication said "I fail to see how 'Silverlight is equally cross-platform as is Flex?' MSFT has announced Windows and Mac only, and are asking a third party to trail their implementation with a lagging Linux port." Speaking of lagging Linux ports, Adobe's Linux port is effectively DOA at this time.
The same representative said "We do have an active developer program…in fact FlashCamp is one aspect of it. Please check out http://www.adobe.com/devnet/ for the many developer tracks, just as a start…there are MANY resources available online and in various forums, both through us and through our community... We are learning how to get developer involvement through a very active focus on Flex itself and getting submissions there is an ongoing goal, unlike other vendors." I again suggest that Adobe look to other software vendors, such as Autodesk, and discover what a developer program actually looks like. Yes, Adobe has released some cool software, but it is no substitute for a comprehensive developer program.
A very senior Adobe employee told me that Tour de Flex was a marketing tool. He may be surprised to know that an individual won't spend the time to run software unless they are well into the sales cycle. Wikipedia has a good definition of marketing. Adobe should read it, and learn how to interpret it for the software development market.
Another spokesperson said "Although you mention that you got access to a wide range of Adobe employees, you don’t seem to find that unusual… I hope your final note also calls out the active involvement of Adobeans with the community, and that we are actively listening." I do not find that listening to customers is unusual for most vendors. Over the years I have had ready access to a wide range of personnel from software vendors such as IBM, Sun, Autodesk, and many others. The answers one will hear are a function of the venue in which questions are asked. Unless one is careful, one will tend to hear the information that one is listening for, in support of one's preconceptions.

Tim Buntel, Senior Product Manager for Flash Builder, posted an article about what's new in Flash Builder 4 at ria.dzone.com. It's telling that all of the reader comments are not about the new product features. Instead, the comments focus on Adobe lack of understanding of developers and the lack of Linux support. The developers that participate in DZone are hard-core. Adobe needs to listen to them, and other developer communities like Slashdot.

_______________________________

Mike Slinn
Independent full-service software contractor and author
http://slinnbooks.com
http://mslinn.com

Comment from Tom Chiverton on 2009-06-18

2009-06-18T15:12:58Z

And to add weight to your thoughts, the *most voted for* bug in the Adobe bug tracker, by a factor of 3 is for just that - Flex Builder on Linux.
That's more important to developers than, for instance, a code formatter or HTML templates.

But don't be disheatened, make sure everyone who cares has voted on http://bugs.adobe.com/jira/browse/FB-19053

There's no comment yet from Adobe on what is going on with this product, but it can't be long before they announce it's going to be updated and maintained, can it ?

Comment from Reto Kiefer on 2009-06-18

2009-06-18T15:23:50Z

One of the best articles I ever read about Adobe and their relationship to a developer community. This attitude has to change if they want to win the developers for their technologies. We develop Flex for more than four years now and are more and more sick about their sub-optimal tools and technologies and their attitude to filed Bugs in the Flex SDK.

Comment from Chris C on 2009-06-18

2009-06-18T16:02:10Z

Agreed its a good article and I can understand exactly where it is coming from as a developer having worked in Microsoft world as well as Adobe world.

However, doesn't it miss out one vital fact - namely that a majority of organisations which currently use Flex are design led and NOT developer led.

That says to me that the design team are always going to outweigh the development team when the choice of development tools is made - I can't see why a design led company using all Adobe tools would suddenly switch to something else when doing development ?

So all in all I guess i'm not suprised by Adobe's attitude to the developer community - just simple economics. In addition - to me anyway things do seem to be getting a bit better with Flash Builder although I'd be happy to whinge all day about compile times.

Comment from Mike Slinn on 2009-06-18

2009-06-18T17:43:19Z

I urge Adobe to make a formal statement right away clearly committing to Linux support, and to publicize it widely. JBuilder's fall from greatness was greatly accelerated because Borland did not make timely commitments to the product once developers started comparing it to Eclipse. Developers are extremely critical of waffling. Nothing less than a strong and clear message of support will be interpreted in a positive manner.

Comment from Jens Wegar on 2009-06-18

2009-06-18T17:52:12Z

Although I do agree on many aspects of the article, in defense of Adobe I do think they have improved on things over the years. Their open source initiatives have grown and the tools have in general always become better than the previous generation.

I personally do not develop on a Linux box, but I can appreciate that many do (or would want to). Adobe should put development of their tools for Linux on the same level with the rest of the Major OS's they support, i.e. new releases should be made public to all platforms at the same time. I do however believe that this is where they're heading, just hope they put that in the fast track. Another thing I agree on is to open up the developer tools for the community to improve on, perhaps only to the extent that only Adobe may approve and publish fixes/patches/updates etc., but all developers could take part in creating new functionality or fixing old, suggest fixes to bugs etc.

Adobe not gaining revenue directly from sales of their developer products, but instead indirectly from increased sales of their other tools is an interesting thought. But then I guess the question is where to draw the line for developer tools? Is the Flash IDE or DreamWeaver also included in this? They're also developer tools, but perhaps more geared towards people from a designer background. How about Flash Catalyst? Or would this idea only include the Flash Builder?

Comment from Chris Gross on 2009-06-18

2009-06-18T18:27:08Z

Most Linux guys are also Open Web advocates. Even if Flash Builder were open sourced, they wouldn't develop for Flash because the Flash player is still closed. If Adobe wants to make a bold move to engage Linux running, free software geeks, they should open the player. At least that wouldn't give up one direct revenue stream.

Comment from Jim Priest on 2009-06-18

2009-06-18T19:25:02Z

Great article! I do hope Adobe wakes up and embraces Linux a bit more.

Comment from Ross R on 2009-06-18

2009-06-18T19:27:34Z

Very good article.

Though, I must point out the growing number of people like myself in the Adobe crowd. People that don't just "design" or "develop" but instead do both. People with a visual background who also go into the trenches and hardcore code with the rest of you. I believe the popular term is "devigner" This population gives Adobe (and others) the best of both worlds. They are people who will update quickly, purchasing hardware and software as it becomes available. But they also play a role in the open source areas.

Microsoft noted this trend first, and attacked it with Expression Blend (though silverlight was still a 4-letter word at that point), and now Adobe is attacking it with the Catalyst/Flash Builder combination. My suggestion would be for Adobe to go the extra step (as it sounds like Microsoft already has) and merge Catalyst and Builder into one application, and then position it as a PART OF the Creative Suite, with full round-tripping, to really tap into that designer crowd (not to mention help out the developers when working with the designer crowd)

Though, there is one thing that Adobe has done that Microsoft hasn't (at least not to my knowledge, and Adobe has not advertised this very well at all). They made Flex Builder 3 and Coldfusion 8 FREE for educators (that means, students, faculty, staff || oh yeah, also for unemployed developers as well in this economy) and they have stated in releases that Flash Builder 4 and Coldfusion 9 will also head that route. (hopefully Catalyst will as well)

On another note, this article, and all those linked, state that silverlight is gaining ground, but only mention the technical-side of the plug-in. How is silverlight doing in actual development. I haven't seen any new major companies working with Silverlight (in fact, now that MLB jumped ship to flash.........). But, I would be interested to see a direct comparison of the actual usage of silverlight vs flash (I would relate it to the amount of time on the market to not give flash an overwhelming head-start) I would also like to see a number-of-developers statistic for flex and flash. Just because silverlight is catching up technically doesn't mean that people are using it.

Comment from Tim Buntel on 2009-06-18

2009-06-18T20:34:50Z

Hi Mike,
Lots of points here, but let me try to comment on the availability of Flash Builder on Linux at least. In March, we conducted a survey of over 1,000 Flex developers and about 7% reported using Linux as their development OS. Meaningless, you say, since there is no Flex development tool for Linux! Whereas the Eclipse.org 2009 Community Survey (promoted to individuals who visited the eclipse.org home page over a one month period around the same time as our survey) shows Linux use on the desktop at 27%. I agree that Linux use as a primary desktop operating system is growing – no debate there. But we face some challenges with Flash Builder on Linux that makes the decision a bit more complex than a normal OS support issue.
I use to work on ColdFusion and adding support for a new operating system (we supported many!) was primarily a QA task. Yes, there might be some engineering details – alternatives for native libraries and such – but CF is a Java application and so it should just run wherever the J2EE app server runs. Just test it and go.
But Flash Builder is Eclipse, you say. Shouldn’t it, like CF on the Java server, run wherever Eclipse runs? Unfortunately it’s not the case. A number of features of Flash Builder do not currently work on Linux. As the Flash platform evolves, so does the need for Flash Builder to integrate with other Adobe products: we have workflows with Catalyst, Flash Professional, and other Adobe products, some custom rendering technologies that make Builder’s design view possible, and a host of other internal parts. It’s not impossible; but making Flash Builder for Linux would require additional engineering work to support. And anytime there’s a question of investing actual engineering work, we need to weigh that feature against all of the other development priorities and features of the product. I’m sure you’re very familiar with this from your time at Borland: each and every feature must be evaluated. Does this feature support the goals of the release better than another feature?
Linux has smaller adoption than Windows, but it’s growing and growing particularly among (as you point out) key influencers and a developer segment who is important to Flex’s success. So should we put our development resources into Linux support, or something else? There’s a large PHP developer population, for example, who are important to Flex’s success too. They use Linux far less, but need other features – particularly around data access and productivity. Which group should be the focus of our efforts on this release? Are Linux users more important to the long term success of the platform than PHP users? My feeling is that they both are important – albeit for different reasons. And what about all of the features that are useful to developers regardless of their OS? Getter/setter generation and network monitor and unit test support, for example, were all features that were highly requested by developers on all platforms. If one engineer works on Linux, all of his effort benefits only those 27% who use Linux (going with the high adoption #) – whereas if he works on unit testing, his efforts benefit developers on *all* of our supported platforms. I am not saying that Linux support is less important than anything else. But in asking the question of what features best support the goals of *this* release, we have had to make some choices based on the available resources on the team and we have not yet decided that Linux support is on the list.
So hire more engineers, you say, and support both Linux and PHP! Give us unit testing and all the rest! Well, we have grown the team – but there’s always more on the list than can be built on time and with the given resources.
I have spoken to many developers who want to see Flash Builder on Linux, but then they say things like “I don’t need all that fluffy stuff – design view and data wizards and Flash Pro integration. I want a good solid coding environment for working with Flex.” But Flash Builder is evolving (as even the recent name change suggests) to be more than just a hard-code coder’s tool for Flex development. It’s about supporting the whole Flash platform for many different kinds of projects and developers. So perhaps what you really need is something different? Just popping that up as a topic for conversation…
Let me close by saying that this is an issue that has driven many lively debates within the team (ahem) – and it’s not over. We have *never* said that we’re killing or otherwise not investing in Linux. But I agree we’ve been too quiet with our plans. We need to let the community know where we are taking our Linux support next, and we need to communicate it clearly. I’ll do my best to make that happen.

Tim Buntel
Sr. Product Manager, Adobe Systems Inc.

Comment from Ryan Stewart on 2009-06-19

2009-06-19T07:19:55Z

Let me start off by saying this is a great article. I think it's great to see someone speak up for Linux and in general for developers at Adobe. We're working our asses off to appeal to developers and the success of the Flex Framework has given us a huge influx of developers with different ideas that we're trying to incorporate. Tim did a great job addressing some of those. And you should know that guys like James Ward are highly respected and screaming for Linux support for Builder. So I appreciate the blog posting and I think it was a great post on developers and Adobe.

I do however, take issue with the "Adobe Responds" section. Not necessarily because you posted it, I don't think any of that was meant to be private, but that you did so in such a way that took snippets and bites followed by your "rebuttal". Frankly, while there was some value in your responses, they ended up coming across as contrite and I think the feedback would have been much better off going directly towards those people instead of making its way into a blog post. I came across as very unprofessional to me.

=Ryan
ryan@adobe.com

Comment from Ryan Stewart on 2009-06-19

2009-06-19T07:38:14Z

Hah. It should say "It came across as very unprofessional to me."

I love the irony though :)

=Ryan
ryan@adobe.com

Comment from Jan Halfar on 2009-06-19

2009-06-19T09:03:37Z

The majority of the (computer science) students working for our company are running Linux on their personal Laptops. The Linux / MacOs / Windows ratio is about 70% / 20% / 10% with a strong trend to Linux.

We are located in Munich, Germany.

Comment from Travis Almand on 2009-06-19

2009-06-19T14:04:41Z

I have mixed feelings on this one. For the most part I think this is a good article and brings many excellent points about RIA's in general.

But this article is all over the place and is not very clear on what exactly the author is trying to get across to the reader.

I do have problems with some of the topics...

Adobe must develop or direct more of their resources towards Linux or they become irrelevant to developers? Why should they when Linux commands such a small percentage of the market? Yes, a decent percentage of developers may use Linux but is it an overwhelming enough amount for Adobe to focus on it without it being a useless money pit? Wouldn't they be better off letting Linux people do what they almost always end doing anyway, building it themselves?

Plus he wants Adobe to invest resources into creating tools aimed at developers needs/wants but admits that developers are reluctant when it comes to actually paying for the tools they use? Again, isn't Adobe better off letting the developers build their own tools which the author admits the developers would rather do anyway?

Yes, of course Silverlight is gaining ground, a plugin that starts with an installed base of nothing has nowhere to go but up. When it reaches 98% user installs then we've reached something. But the article linked to with this statement isn't even about install numbers but Silverlight features as compared to Flash. Silverlight is gaining ground in trying to implement features Flash already has.

And to the end user, Silverlight market share versus Flash market share is irrelevant. Why? Because I can have both installed at the same time. Any statements about market share of Silverlight as compared to Flash is pointless. If I install Silverlight it's not as if I have to uninstall Flash. The end user in most cases will not care if it is Silverlight or Flash, they just want it to work.

The number one complaint about Flex is lack of proper Linux support. Ok, I understand that one. But we are given an example of good developer support with Microsoft? How much has Microsoft devoted to porting their .Net and Silverlight development tools to Linux? I'm really curious because I've never heard of such a thing. All I've seen is that they are more or less not standing in the way of preventing third-parties from developing these tools themselves. Not to mention a lack of support for OS X with MS development tools which that group is very vocal about.

I have to say I'm really tired of people suggesting Silverlight might be superior to Flash because it is so easy for .Net developers to move over to it. Of course it is, it's designed that way. It uses existing Microsoft development tools in much the same way as anyone who's used the tools before for .Net development. There is no such comparison with Flash/Flex. How about we focus more on people who have never developed at all and are just starting out with their careers? Or how about people start mentioning that Javascript developers probably would have an easier time to moving to Actionscript development because they are almost the same language?

Flex Builder is not a mature product because it doesn't compile as fast as he wants as compared to a similar Java project? Huh? In that case Java is not very mature because it doesn't compile as fast as similar Javascript projects. Ok, that's a stretch but hopefully you get what I mean by that.

Adobe is behind the game when it comes to supporting developers but not because of lack of supporting Linux. For years Adobe has made products for designers and I imagine it is very tough for them to get into the mindset of developers, much like the author describes. One of their solutions to this was to fracture the Flash platform to a terrible degree. There is absolutely no reason whatsoever (actually it's ridiculous) for there to be so many products aimed at one thing, Flash platform development. It would as if there were multiple versions of Photoshop; one for png, one for jpeg, one for gif in an effort to support web graphics.

There is one excellent thing about Silverlight, it provides proper competition to Flash/Flex which in the end will make both platforms better.

Comment from Patrick Whittingham on 2009-06-19

2009-06-19T15:34:29Z

Over 90% of our IT professionals are Windows at a large US company.

Comment from Craig Kaminsky on 2009-06-19

2009-06-19T17:35:25Z

This is an excellent and thought-provoking article. I develop on a Mac and went through the 'pain' of waiting for many pieces of software I used to use on Windows to make their way over to the Mac. I'm nearing the point of getting a new laptop and am strongly considering a Linux OS. The lack of some software for Linux does hold me back (i.e., Flash Builder :).

However, I don't feel that Adobe has a poor developer program. I just think it's still evolving. This is a company that made it's success on the back of printers and design applications. It's only quite recent that they've delved into the server world and that, really, was only due to the Macromedia purchase (picking up the Flex 1/1.5 (or was it v2?) server and CF 7 from Macromedia). Adobe has only released one version of CF as their own and, unlike Macromedia, is truly working on an IDE for CFML (sorry, HomeSite fans, I wasn't one!). As a result, it's good to let them know what we (developers) need but be careful not to rush to judgement. Just as it's inappropriate to evaluate a professional athlete's entire career after 1-2 years, so too is it inappropriate to give a final pronouncement on Adobe's developer programs, such that they are (at times, the article reads this way to me).

One final thought on Adobe and Linux: personally, I cannot fault a company for making a smart financial decision. Seriously, in my limited involvement in Linux world (and I readily admit it's new to me), no one pays for anything. Everything (software) I've used, as well as those Linux developers I know, it's all open-source. So, I'm not so sure about faulting Adobe for not jumping on the Linux bandwagon. Will you true Linux developers really spend $300-$400 for an IDE when you could use the SDK and your 'normal' IDEs to build the same thing?

Comment from Mike Slinn on 2009-06-20

2009-06-20T20:57:33Z

I believe this is the "most voted for bug" that Tom Chiverton made reference to: https://bugs.adobe.com/jira/browse/FB-19053

Please make your vote heard.

Comment from Joeflash on 2009-06-21

2009-06-22T01:30:53Z

If numbers of developers was the deciding factor for tech adoption, we'd all be building applets for the Java browser runtime instead of for the Flash Player, because there were so many Java developers ten years ago. Truth is, the reason Flash has succeeded IMO comes down to two things:
a) the designer tools for Flash enabling the creation of compelling user experiences;
b) Flash video catapulted the technology into a mass usage in 2003, and it was at this point when a critical mass of Flash Player installs was achieved.

Businesses building for the web generally look for the technology which will allow for the widest audience and the most capable expression of the business case, i.e. the user experience. The Flash Platform has not only been a leader in this area, it has literally redefined what is possible on the web. First adopters were design agencies and heavily branded and multimedia-driven experiences. Second adopters of the Flash Platform are enterprise users, and yes this is where Adobe does need more work than perhaps Microsoft on the developer side of things.

But if RIA development has one thing in common no matter which technology you are talking about, its that this is a medium that necessitates a complete workflow between both designers and developers, leveraging both kinds of creators, which is something many developers in my view have yet to wrap their heads around. If you've ever used Microsoft's creative tools, you'll know they don't hold a candle to Adobe's tools. So it's all relative. Microsoft would best adopt a strategy that leverages other company's creative tools, and focus at what they're good at, and Adobe would best increase its moderate offerings on the developer side of things while leveraging other technologies, which it has done by investing in the Eclipse platform.

If you want to play in the RIA space, you've got to cater to both kinds of creators, and I personally think Adobe is doing a pretty decent job at this, even though I agree, Adobe would do well to increase its marketing spending and outreach in traditional developer circles. ActionScript is still not accepted as a serious developer language, despite the fact that it's just as capable than some of the more established developer languages such as Java, .Net or Ruby.

Speaking of closed technology, I personally am tired of explaining to non-Flash Platform developers why the Flash Player not being open is a very very good thing, why it is nothing less than the critical pillar of strength upon which the entire history and future of the technology resides. The problem with browser based technologies -- and why all the noise about HTML 5 is sounding rather vacuous to me right now -- is that there is no unified "runtime" (i.e. the browser). If different open source communities had their own version of the Flash runtime, the Flash Platform would be no different than traditional web-based development, prone to all the same inconsistencies and lack of universal implementation support. AJAX tries to make work around this weakness in traditional web development, but let's face it, it's a hack, plain and simple. It should not take a billion-dollar company (Google) with a team of uber-engineers just to design a ground-breaking interactive application. Google Wave looks nice, but man, that is way more work than most of us mere mortals are prepared to expend on an app, including enterprise companies.

And the idea that because the technology is closed, it is vulnerable to Adobe "someday abandoning Flash" is the most absolutely ridiculous thing I have ever heard of. It would be like Microsoft abandoning the Windows line, or Google getting out of the SE business -- ain't gonna happen, EVER.

So yeah (like for the millionth time I seem to be saying), the Flash Player is closed (and that is a GOOD thing), but that does not mean that the Flash Platform is a closed community. Far from it.

Comment from Joeflash on 2009-06-21

2009-06-22T01:34:24Z

Sorry, preface that last comment with the following:

Tom Chiverton said "the *most voted for* bug in the Adobe bug tracker, by a factor of 3 is for just that - Flex Builder on Linux." Actually that's not entirely correct - a very close second is bug FP-444, "Ability to intercept system error dialogs," which is another big hole in Adobe's technology chain that needs to be fixed ASAP if developers are to treat the technology as a serious contender for enterprise development.

I am in agreement with what Travis Almand said about being "really tired of people suggesting Silverlight might be superior to Flash because it is so easy for .Net developers to move over to it." LOL, I agree. Adobe can claim that Flash Builder is superior to Silverlight because it leverages all of the Flash ActionScript developers who previously were forced into using an inadequate IDE for coding, and all the JavaScript developers, who don't have to learn a new language to port over to ActionScript (seen as the two were based on the same ECMAScript standard until recently), and leverages all the Java developers because ActionScript is very similar to Java and uses an Eclipse based IDE that they are very familiar with. So the argument of "who leverages existing developer communities better" is meaningless, because each technology has a following, and a history, and number of developers using a technology alone will not win you the day, although it is an important factor.

If numbers of developers was the deciding factor for tech adoption, ... (see prev post)

Comment from Tom Chiverton on 2009-06-22

2009-06-22T08:20:36Z

@Joeflash: I should have said 'in Flex Builder' when talking about bugs, you're right.

Comment from Joeflash on 2009-06-22

2009-06-22T20:53:40Z

When I said, "Adobe can claim that Flash Builder is superior to Silverlight..." I meant Flash & Silverlight.

Comment from Andrew Scott on 2009-06-24

2009-06-24T11:38:25Z

Well we claim that hisotry never repeats itself:-)

Yet Adobe haven't learned much from their own past. The one thing that Adobe could learn from are 2 of their most flagship products PhotoShop and Illustrator, both of which in the mid 90's were also focused an lead by designers and not coders.

Both of those Applications were years behind in functionality, and features on the Windows platform. They were more focused on the need of the designer and their core market than they were ever interested in, and ColdFusion and thier Eclipse based products are now suffering that same fate.

I hate to dredge this up, but on one hand ColdFusion has moved in some areas where it could not have hoped for if Adobe didn't buy it. But on the other hand it has suffered the lack of understanding that Adobe should have learnt from back in the early to mid 90's

Adobe has only proven that history can rpeat itself, and that they can only and will only focus on their core market, or should I say they are blind sighted to what they think is their core market and refuse to acknowledge that there is more potential that could be tapped into.

Even from a design and coding point of view, Flash Builder is still not to the standards I would have though a Design and Coding IDE would be for such a product and by that I mean the application relies to much on manual coding and very little on the design and UI aspects of the IDE to help get the most basic things happening quicker.

Comment from Ross R on 2009-06-24

2009-06-24T13:33:04Z

@Andrew:

I agree - though I thought the saying was "History always repeats itself"

Comment from Todd on 2009-06-24

2009-06-24T21:53:49Z

Maybe what the open source Linux guys need are better EMACS extensions, since that's what they'd use anyway...or is VIM the hotness these days?

I don't blame Adobe for not throwing money after a problem that's going to bring in NO foreseeable revenue. If you really need to use Flex Builder and you have a client throwing money at you, go grab a $999 Mac or a Windows license for $50.

Until then, there are guys that are able to get by just fine with the FREE SDK and the command line tools.

Comment from Michael Borbor on 2009-06-24

2009-06-25T01:30:51Z

I think you confused something in your article, it's the Flex SDK not the Flash SDK.

Comment from Keith Sutton on 2009-06-26

2009-06-26T17:13:59Z

Like your rant by the way and am compiling my own post on the subject of developers as I see 2 subgroups of developers:

1 Flash Developers that have a history/lean to Flash as they have come from the Design-side
2 Business/Enterprise Developers that have come from Java/MS/Oracle world and build traditional applications with very little 'Design' influence.

I singled myself over disagreeing with the Flex-to-Flash Builder re-naming and although this did not go over well with many Adobites that are predominantly Design/Flash leaning (SubGroup 1). I contend that the name change is not only compromising the credibility of SubGroup 2 (simply stated they cannot put 'Flash Builder' on their resume because they will not be perceived as application builders by their business/enterprise customers) but also 2-3 years of hard earned respect that Flex Developers and achieved for Flex RIA business applications. Since airing my view I have been approached by many in the Startup/Business/Enterprise application SubGroup 2 who agree and are very worried about 'loosing the RIA war'.

I like you idea about serving the developer community in a strategic manner, particularly Linux-based community which, I believe, correlates to Group 2 and an enabler to driving sales of LC and other enterprise offerings beyond PDF and workflow. Separating out suites for Design, Flash Developers (1) and Business Application Developers (2) would serve the community and business opportunities better.

Comment from Raja Patil on 2009-09-12

2009-09-12T12:47:43Z

I also strongly supporting opinion expressed in the post about
having flash builder on linux. For last 5 to 6 years I am looking
for good GUI development tool on linux. I tried Kylix also
but even today I am not finding any good tool like deplhi
for GUI development on linux. I started learning Flex because
its a good GUI tool and its available on linux too. But as of today
Flex/Flash is still not supporting Linux. I am sad about it.

There are many like me around who are forced to use Windows
for development though wish to embrace linux but could not because of fact mentioned above. The author has expressed
the pain exactly. Many like me who are bogged up with daily
work pressures do not express themselves in the manner the author has expressed but it does not mean that potential is not there.
Adobe should consider this and can invest on linux. I am sure it will pay them off in terms of revenue due to the hidden unexpressed potential.

Hope Adobe takes me as representative of non-expressive developer community who rarely air out concern.

Thanks.

Comment from TK on 2009-10-09

2009-10-09T22:27:59Z

I wholeheartedly agree. We developers are a different breed. Java developers can develop crossplatform using JDT, so why not Flex? Open sourcing Flash Builder would push Flash Platform RIA's into areas unseen, and break any records achievable with a closed dev. platform. Help us help you, Adobe.

Comment from Bruno Fonzi on 2009-10-21

2009-10-21T16:31:42Z

Powerflasher, the creators of the FDT, the Flash/Flex IDE, is seriously considering investing to build a Linux version, if interested let us know.