Speaking about the Presentation Layer, I still see most of our clients using Struts, and a few of them using JSF. In France, I only saw one company using Flex. I'd say Jeff Whatcott might be too optimistic, especially because Flex is not open-source and we -Java users- live an open-source world...

Why Flex breaks the Web

Why Flex breaks the Web

Hi Shashank,

I think Jeff is being conservative (that said, Jeff and I work together at Adobe, so no lack of bias here -;). Thanks to network effects, the adoption curve for a new technology is exponential, not linear. (eg. It is much easier and faster to go from 1/2 a million to a million than from zero to 50,000.) We've had something like 150,000 folks download and try Flex since we launched the Flex 2 public beta. We have an active community building real apps of tens of thousands today. (There are over 5000 I believe now on a single mailing list, flexcoders). We have a lot of very high profile apps coming. The release of Tamarin last year, Apollo this year, some upcoming new technology, 6 new books by O'Reilly alone, will all draw in new folks.

The best sign for me that viral organic uptake to get us on that exponential curve is happening is a conference called Flex360. This is a full multiday conference on Flex and it was organized and hosted by eBay, not by us! The conference sold out 400 spots six weeks in advance of the event. The day it sold out, by chance, a company in Washington DC announced their own Flex conference for June. This stuff is happening now more and more around the world. (We had 2000 people at an event for Flex recently in Korea!)

I think we'll exceed these expectations.

What do you think of Flex yourself? If you haven't tried it, this is a good place to start if you are coming from a Java background: http://coenraets.org/blog/2007/01/flex-test-drive-server-for-java-developers-tomcat-based/

Regards,
David
Adobe

Shashank, interesting blog post, but this is also preposterous. Flex has very limited market penetration, and almost every Java developer (and some Flash programmers) that I've spoken to, don't consider Flex to be anything more than a very limited use proprietary framework.

Flash is king when it comes to rich clients on the web, it can certainly trump AJAX, but Flex. I'd say it is more like 1 in 100 Java developers is learning Flex today.

Tim, I am only wondering if what Adobe is claiming, or should I say aiming for, has any rational basis.

Shashank, it's called marketing hype. I'd say that there is no rational basis for what they are claiming. It's that or they are talking to the wrong market research provider.

The problem starts with statements like "We have somewhere between 3 and 4 million Java developers". What does that even mean? What's the margin of error? Who asked the question even? I was at a client's office a few weeks back and I saw a headline from SDTimes - "World to see 12 million developers". To me it's unscientific extrapolation.

David's comment: "The best sign for me that viral organic uptake to get us on that exponential curve is happening is a conference called Flex360" OK, # of downloads not terribly important, anyone will download your software, a few will install it, and of those that install it even less will use it.

To me, the important # in David's comment is 5000. There are 5000 watching the mailing list. And, of this there is a subset actually using Flex. The number is more likely to be closer to 50 or 100 thousand than 1/2 million.

Idealistic extrapolation on top of optimistic marketing. Flex has promise, but don't try to convince us it's caught on before it really has.

Yeah, I don't buy it either. I think 1m people might have toyed with Flex, but actually deploying applications? I don't see it happening. I don't have anything against Flex. Quite the opposite, I think it is a pretty cool product. However, the original licensing model was just way WAY out of bounds and the new one isn't much (if any) better, at least in terms of adoption in the enterprise space.

These numbers are not unlike "counting downloads." Downloads of your product don't map 1:1 or likely even 5 or 6:1 with "people", certainly not with "in use".

Hi,

Let me try again...I hate to be accused of "hype" -;)

We are not claiming we have a 1/2 million developers. That is an extrapoloation that blog author made. I disagree with that. My assertion (based on a bunch of data sources) is that we have about 150K who have downloaded or tried Flex and 10s of thousands active developers. That is not 1/2 a million, nor a million, but it is a great start for a product that we shipped 7 months ago and I believe a sufficient start (but not a guarantee) for us to hit the goals that Jeff suggested in the article. This is not all marketing "hype". There is a pretty deep body of data on the developer community, # of developers per region, by language, by application type, etc, from companies like Evans Research based on statistically valid surveys of developers around the world. Here at Adobe (and those of us who come from Macromedia, or even Allaire before that) have a lot of experience building such communities. We have an active community of about 300K ColdFusion developers, well over a million Flash designer/developers, and several million Dreamweaver customers as well as folks on the team that led the creation of the .NET framework and key parts of Java/JSP. We've tracked these kinds of trends and adoption rates before. That said, past results give us insight into how an adoption curve *could occur* to meet our goals, but the rubber will hit the road on how well people like Flex and whether it solves real problems for them. So far, the trends are good (from our perspective) but the real question I have for you is what do you think of Flex, rather than what do you think of how successfull Flex is? Time will tell on the later.

Regards,
David

I think Flex is an awesome product and I haven't seen any product that comes near it in terms of GUI power and productivity. Backbase comes closest but those licenses will *really* scare you. I think the license scheme for the Flex Eclipse builder is quite reasonable. And when it comes to open source, most AJAX frameworks are either inflexible or require a lot of expertise on the side of the developer. You get what you pay for.

The current license scheme is a significant improvement although I don't like the huge markup on the data integration. That part of the product is essential for enterprise goals and the seperation seems artificial to me. Otherwise you're stuck with XML and webservices requiring significant server side coding efforts and performance degradation. The license scheme there reminds me of the old Flex 1.5 license scheme again.

The biggest gripe I have with Flex is startup performance. It takes quite a bit of time to download the base libraries. These are then cached, but still.

Forget the tiered licensing model as a decision maker for Flex. Fist evaluate this product and make sure that your Java development can be tied to Flex (and more specifically Flash and ActionScripting). We had to abondon Flex because of too many mismatches with Java objects and ActionScript with the Flex gateway. We tried to force development by augmenting our code to fit with Flex. Our main issues involved trying to match Java objects (POJOs) with ActionScript objects with Hibernate and dynamic proxy handling. Allowing proxy objects to pass through the gateway and maintaining relationships was too much for us. It also helped to have a Flash developer on staff to help the Java team understand the details of what Flex is doing under the covers Just make sure you evaluate the framework capabilities before buying this hype.

I'm an Adobe employee working on the Flex Data Service project. I wanted to respond quickly to Mark's comment on using Flex with Hibernate. We do have technology in FDS 2.0 which allows flex clients to use and manipulate hibernate data models. It deals with hibernate proxies by prefetching any data required while the transaction is still open and then unwrapping proxies during serialization. It also can be used to automatically update hibernate object models. It supports such features as lazy loading of references, paging of large collections, and conflict detection and resolution. The source to the hibernate adapter is shipped with the product so you can see how we solved these problems. Additionally, there are a couple of open source projects which will automatically generate strongly typed action script objects from your Java domain model using xdoclet. Things are moving pretty quickly in the Flex/Java integration space as we at Adobe are very focused on making Flex easy to use for Java programmers.

Yep, count me as another 1 in the 10

Flex, along with other Flash-based solutions like Laszlo, sucks because Flash is not an integrated part of any OS or browser. A Flash app running in a browser window is an even more ridiculous thing than a crippled 32-bit application running on Windows3.1 through win32s thunk. The splotches of Flash UI, which appear here and there in increasing numbers, do not react properly to keyboard, have fixed size, have fixed-sized fonts, have non-standard widgets, cannot be nicely printed (though I've heard that ActionScript supports CSS, so targeting print media should not be a problem in theory), are not integrated with other elements of a page or even with another Flash splotches on the same page, the UI itself is painstakingly slow and juddery, Flash objects are reloaded every time you go forward and back (even Google Finance does that, when you return to a page the stock value graph does not recover the parameters you set before leaving the page).

Leave Flash to what it does best -- annoying ads and video clips.

1.) With Flex 2 the server side has not to be one bit of Java until you use Flex Data Services. You have SWF´s for the client and you can use any server side technology that supports the HTTP protocol or SOAP-based Web Services.

2.) I don´t want to say anything to the mentioned numbers in the article. I think the upcoming desktop runtime "Apollo" will significantly increase the impact on the flash/flex technology. I can´t imagine that nearly any AJAX developer wants to ignore the possibility to port his/her application to the desktop and sooner or later they are going to use the one or other flash technology (Drawing-API, Videos) because the flash-runtime is there, if they use it or not.
Of course this potential transfer is vice versa but I can´t think of one AJAX tool that a flash-developer is intensive looking for if he/she can use the powerful flash- and flex-api´s.
Instead complete websites or webapps are embeddable with the including html-runtime.

3.) There is no "Flash OR Flex". You can combine your work from these two IDE´s perfectly. After all the end products are the same: SFW´s.
Flash stays the overall number one for the flash-platform but Flex is extremely powerful when it comes to RAD (and it gives the one or other ignorant Java/.NET developer another try).

4.) To Michael ISVY:
It´s right that Flex is not completely open source like Java with the emphasis on "completely":

- the class library is open source. You can look at it and extend it your way.
- the SDK (compiler) is free and NOT open source. But show me one Java developer who is glad about the need to target multiple (different) Java runtimes from multiple vendors!

5.) To Michael J.: Sorry that I have to say it this straight way but your comment is another case of "I don´t know the technology but I love to talk about." or simply "flash-misunderstood":

- "Flash-based solutions like Laszlo, sucks because Flash is not an integrated part of any OS or browser".

Alright. Flash sucks because it is OS-platform independent (like and better than Java) and mostly pre-installed on Windows, Mac and more and more Handhelds!?

- "... do not react properly to keyboard, have fixed size, have fixed-sized fonts, have non-standard widgets, cannot be nicely printed ... does not recover the parameters you set before leaving the page)."

Yea, and Java does not support I/O pretty well, right?
You have to be careful with the distinction of what a technology doesn´t support and what a developer doesn´t use!
Nearly everything you wrote is wrong.

- "Leave Flash to what it does best -- annoying ads and video clips."

And leave Java to what it does best -- server side work.

I use openLaszlo to build an application for my customer and am quite happy with that. Adobe Flex is too expensive for me.

Adobe Flex is too expensive for me.

The Flex SDK is free. Did you try the free Flex SDK?

Let the best technology prevail!

I just want to remind everyone that eventually, we ARE going to have a STAR TREK universal programming language called English in which you simply talk to a computer and it understands prefectly what you WANT it to do instead of what you TOLD it to do ;)

Until such time, learning a multitude of technologies is the Programmer's job, this is what distinguishes between programmers, the one who knows as many technologies as is humanly possible should in theory be better at what he/she does. The success of one technology does not have to come over other technologies as sometimes you will see software solutions that require a mix of several technologies at once, sometimes even rival ones.

We will use Flex in a coming project. Most of the UI will adopt Ajex, and a few complex UI will implement using Flex.

Hi guys, I am head of software architecture in my company and we have evaluated several frameworks like OpenLazlo, Flex and some Ajax stuff as well.
I think it is ridiculous to think that 1 million developers will just start developing in Flex. The Flex development suite is not free at all while on the other side, there are plenty of free-to use Ajax frameworks.
"Flex SDK is free" is like telling the Java SDK is free, who cares, I mean not any single developer is using javac on the command line anymore!
The nicest Eclipse based tools (Flex builder) which can really help starting to develop in Flex are not free.
I think Adobe has a coherent vision, Apollo is promising, the Flash runtime is ubiquitus but my personal opinion is that if Adobe wants Flex to become popular as well, they MUST give away the Flex builder as well.
Robin

3 out of 10 statistical statement are made up.

All the power to Flex and rich media. I don't understand why this scares some people in the industry. The more UI tools the better. Let's face it AJAX is nothing but a way around the browsers limitations and if Flex can provide a usable and "developable" GUI, why not use it?

Most critics of it say something like it doesn't have print or no back button but those problems are solvable. Once high connection speeds become a norm, people will laugh at AJAX.

Dear Adobe,
I'm a designer and easily hooked. So please give me a free and easy (think paint-by-numbers) environment to play around with. Once I get a good feel for it, I can start creating some amazing designs. If you can convince the tech guys that it's easy to deploy & maintain across platforms, and the business guys that it won't cost them a leg and an arm, we'll all be happy ever after. Thank you!

p.s. quote: [snip] we ARE going to have a STAR TREK universal programming language called English [snip] => More likely to be Mandarin/Chinese, actually ;-)

Cmon folks. Flex is the bomb. Time to drink the kool-aid. Most Java developers I've come across, when they actually use it - LIKE IT AND GET IT.

Within Java, there are a ton of AJAXy frameworks and tools out there, that fundamentally were designed to get around the weaknesses of the HTML paradigm.

Most IT managers out there don't what the heck to use. We are stuck in a transition period, a paradigm shift of technologies. We are unsure what technologies are going to survive multi-year cycles of product development, what kind of talent will be available and what the next technology will be that we missed out on that we have to sheepishly report to the CEO. 'Sorry, this years budget will be spent on rewriting EVERYTHING.'

Managers and architects don't know what direction to go in. Java folk in particular have gained tremendously from Open Source and will continue to do so, but sometimes you have to bite the bullet and realize there are things out there that just provide the answer - that may not be OS.

Within Java - what do you use? JSP, JSF, Tapestry, Webwork, Wicket, SpringMVC, GWT...? We had put together a framework using Tapestry, one of the best component based Java frameworks out there. Your typical Tapestry dev is not a lightfoot. Unfortunately Tapestry people are hard to find, it is not backed by a company, there are not adequate tools and it has a steep learning curve. We chose OpenLaszlo as our next generation RIA stack, but after trying Flex we're at a key decision point.

For all the naysayers - just download and try it. FlexBuilder provides the ease of use of original VB with greater power and a resulting nicer UI. It is great even just for wireframing. ;) Within Flex, there is a great widget set, a good event model, e4x is very powerful for manipulating XML as objects (though at times a tad difficult to get used to). With Flex, you have a good combination of OO + scripting in one. On the backend, don't abandon your Java, just accept it's not good for presentation.

Within the Open Source world you're not going to get everything for free. For the functionality it provides, I am willing to plunk down $500/seat for my developers for Flexbuilder. This is the equivalent of 1 billable day for a contract Java developer in the US.

Quite honestly, HTML is still great for documents and websites. But for true applications, Flex is the way to go.

The biggest hurdle was the pricing model, something that Adobe screwed up on originally that hampered people even looking at the product. I remember from the Generator 1.0 days, extra zeroes just turn everybody off.

Something Adobe doesn't want to publicize too much is: YOU DON'T HAVE TO USE FLEX DATA SERVICES and this is where the expensive part comes in. This is the tough dilemna that Adobe is in with promotion. They still need to get folk on the Flex bandwagon and announce 'we've changed our licensing model' but still have to make money. They are fundamentally not a services company - they sell products.

If you are doing real enterprise dev the licensing costs should not be a factor. Otherwise if you already have a Java backend, likely using Hibernate and Spring, you can expose your data in a myriad of ways. We created generation tools that expose it as RESTful services and it is so easy to wrapper via Flex. I believe you can access your Spring beans directly as well though we have yet to dive into this fully.

I don't know about true usage numbers and hype (I wish they would do some more scientific analysis or surveys), but Flex has a larger community than OpenLaszlo in a shorter amount of time (please correct if wrong), it's RIA Flash competitor, which has some great stuff but is essentially losing the battle. I know there will be other sticking points down the road, but they will be surmountable.

Flex + Java is good company. The kool-aid tastes good and is free of poisons. Hop aboard, come down the rabbit hole, there is a new world awaiting you.

I should disclose upfront that I work at Microsoft, but prior to Microsoft I was playing in all three pieces (java, coldfusion and flex). So that being said, is it possible that I can take off my MSFT hat for a second or two? (understandbly the answer maybe no).

I'd say that there would be quite a lot of folks trying FLEX out (I agree, that maybe the numbers are conservative) but the ultimate metric remains on those developers sticking out with FLEX post the buzz period.

AJAX had gained more momentum then most orginally thought it would (myself included) and despite its limitations it is more appealing to the average developer (of all flavours) as depending on the chosen AJAX framework, it does compliment their existing skillsets (no lurning curve).

I'd argue the sustainability of FLEX post its initial download is more the metric of focus.

Scott.
Developer Evangelist,
Microsoft.

I have been working on Flex from past 10 months(am from the java world). I started with Flex 1.5 and now exploring Flex 2 the one thing that hurts is the pricing of the Flex data services 2. I was happy using RemoteService in Flex 1.5 but now with Flex 2 it doesnt come FREE. Am thinking of developing and deploying an RIA application and going commercial, is there anybody who knows how to integrate Flex 2 with Java server side(exchange custom java/AS objects) without using FDS.

Don't worry guys,

You can contribute to this project www.graniteds.org or just wait and see. I'm sure that more projects like this will appear (as happened with OSFlash)

The community is growing faster and that's great for all!

I think Adobe has to release a Flex Builder Express (or something) for free as Microsoft did.

Like Steve Motola said, we are in a transition period.
I place my bet on Flex and I'm trying to "evangelize" developers and managers here in Argentina to adopt it.

Anyway I'm looking very closely at the RIA space....Silverlight, JavaFX,....you'll never know....;)

Am a java developer myself and i know it is relatively easy for java developers to learn flex, but that doesn't make it encouraging.Flex is not even open source plus you have to rely on other technologies to create an enterprise application.For traditional java developers, i would say its 1 in every 500..

It would appear that you're assuming that all the flex developers will come from a Java background, and this assumption is flawed. The main draw of Flex is the ease of developement. I jumped to Flex from Coldfusion. If I knew Java then I probably wouldn't be using Flex as my developement platform because of some of the Flex limitations. However, I'm currently developeing a commercial application using Flex and Zinc having never programmed a line of Actionscript code before starting on this project. Kudos to Adobe for creating a relatively user friendly developement tool for those who have no experience in Java or Actionscript.