For Farata Systems the year 2007 started with a surge of requests for working on new Flex projects, which was expected, but a number of requests fall under umbrella “Save our project“. For consulting firms, it may sound good, because any repair job is more expensive than creation of the product from scratch. But it may also represent a bad trend – some enterprise development managers will fall into the same trap – the project should go to production in a month, but it’s not in a good shape. Let’s see what happened.
Adobe markets Flex as a RAD tool. The main threads of most of their Flex presentations is how to develop a working application under ten minutes or so. But seasoned enterprise developers understand that there is big difference between a short demo and development of a real world application, which most likely is more complex that reading an XML feed into a plain data grid. But what if your organization does not have seasoned Flex developers? What if you did not have a say during the budget planning? What if your project plan was painted in large brush strokes? It does not take a rocket scientist to figure out that you might not meet your deadlines, which were set by some “bad executive”.
If you’ve been around in the industry for a while, you may remember that the forth generation client server tools like SQL Windows (Gupta) and PowerBuilder were the hottest tools on the market about fifteen year ago. Back than, if you knew how to disable a Button and how to write an SQL join (Hibernate was not invented yet), you’d get a job in no time. Then, you’d learn on the go. This was a Client-Server Wild West. Now we are entering the RIA golden rush. This is great, but do not expect Flex to be a silver bullet. It’s a great tool, but you’ll have to deal with same old issues like performance, application size, custom data renderers, dealing with J2EE components, and more.
That’s why we, and I’m sure many other Flex consultancies started to hear S.O.P. (Save Our Project) signals. Flex is a great tool but do not get overly excited, invest some serious money in training of your developers, and carefully plan your transition to the great RIA world, where every serious enterprise is moving.
Anything that can be clicked and dragged and called RAD can be a dangerous thing and must be used with caution. History will definitely repeat itself with folks not understanding the dangers involved in ‘I can simply drag this on to here and I never have to know how it works..”.
I still feel ‘unclean’ every time I think of VB6 and Visual Studio.
I’m not looking to score some “anti-Adobe” points here, because i’ve been in the SOP spot a few times (especially when price was reduced from 15k to zero).
This is the part you don’t read in the glossay brochure, it does hurt and it does take time. Rushing out and adopting FLEX is one thing, surviving it is entirely a different challenge.
I made a post on this exact thing in Oct 2005: http://www.mossyblog.com/archives/539.cfm “Don’t gamble your career with FLEX 2.0″.
Call Microsoft what you will, but coming into the company from Adobe Development, I’ve noticed one thing above all: “It’s much easier to build in their tools then it is Adobe’s”. Adobe are focusing on the output, which is great and understand why – but – they aren’t devoting attention to the input, you the developer and that’s why people keep tripping themselves up with FLEX.
I think you’ve got it backwards, and I think you might find out soon enough . Microsoft business is long past the time of developer’s tools, and Macromedia is primarily toolmaker for developers. As a matter of fact, my personal communications with microsoft developers support and product managers show them as irresponsive and arrogant. I believe that attitude caused loosing market share of IE and failure to evolve AJAX (formerly known as DHTML) into application platform over period of 7 YEARS. I actually personally heard from IE product manager that MSFT is not interested in fixing bugs in low-level functionality affecting performance and robustness of Web applications as they believe the worse it is going to be, the more incentives developers will have to switch to Vista – and that was over 2.5 years ago. Inability to see the obvious and following blindly corporate marching orders does not get you love in development community.
Second, Flex, unlike publishing tools, is geared toward client-centric applications. Client-server old-timers will find themselves comfortable sooner then J2EE veterans. Knowledge transfer is very important unless your company is ok with SOP situation. Recognising potential and inherited risks requires full awareness and information that is hard to get out given the marketing noise level and wrong assessments.
For those of us who programmed over few generations of technologies Flex/FLash platform is a VM for going where computers never went before – with full support of the very talented and motivated Flex team and community. I think that most of the people in this community will feel very strongly about Microsoft claims of being developers oriented. For the last 10 years the only good products for new development coming from MSFT were former employees.
Personally, the attitude of evangelizing microsoft via cheap tricks like ” I’d find the most vocal person who hates us, and pick a fight, and I would do so knowing I would lose” http://www.builderau.com.au/news/soa/MS_Adobe_war_in_blogosphere/0,339028227,339273152,00.htm) greatly diminishes your credibility. It also speaks volumes on MSFT hiring practices for evangelist position. I would spend at least a year doing real application DEVELOPMENT experience with alphas and betas of the products before taking on Evangelist position. You really have to believe in things you are doing to succeed yourself and make other to succeed,
That being said, Flex does let you build components + animation effects faster…however this might not carry as much weight for a corporate app used by traders and wall street wonks.
For new greenfield projects that can consider a desktop client, at the moment building on an RCP is easily the winner.
Anatole: I stumbled upon this post today, so forgive the late response
I’m basing my crit on the fact that I’ve spent countless hours, pain and suffering building with Adobe/Macromedia products. I’m in the luxurious position of having the time to do a comparison on both sets, and that’s why I’ve held off for quite some time in broadcasting the deep dives in the products. As I’ve spent most of the time getting deeper and deeper into the web based tools Microsoft has on offer (2months in) and my original opinion stands, It’s much easier to punch out applications inside MS then it is with Adobe’s coutnerparts. I was disturbed at how easy it was to put together an AJAX ASP.NET application in comparision to Coldfusion, in both IDE and execution of OOP patterns. CF has its plus and minus but overall i’d rate ASP.NET easier to work with (that’s after 9 or so years of working with Coldfusion).
FLEX is painful, and I’ve seen a lot of people state “Flex Hurts” and your post is one of them. I’ll pick fights with folks who hate us, just because I want to get to the bottom of why, and secondly I’d prefer they do so out of their own home grown research and use with “alphas” and “betas” instead of subscribing to the mainstream channels opinion. In the case between the article you quoted, I felt that was the case. Ted’s not a fan of Microsoft for whatever the reason he decides – which is cool, I don’t have anything against that – I only jump in when I feel there is an unfair message being portrayed by a visible ranking Adobe FTE (just like they do with Microsoft and countless other companies that dare strike out against Adobe’s PR).
My hiring was a hard process to get through, 4 months of rounds to be honest. As for questioning my credentials, i’d probably respond with a question around “what is it you actually pioneer in the RIA space”
Microsoft – Sticks n Stones, will actually break bones if you throw them hard enough.
There is a big difference between pioneering things and taking over. In my experience Microsoft made 0 efforts in 6 years (1999-2005) to do pioneering work with AJAX and RIA and a lot to own the space afterwards.
I do not think MSFT should get credit for making AJAX easy to code – that was mostly work by unknown JS developers suffering through these years. However, it is the one to blame for keeping browser unreliable, slow and unsecured/untrusted platform – and that would be their direct responsibility as the platform makers. By keeping browser bad they thought upgrade to Vista would be more compelling. IMHO, they lost all the credit they built in 90s.
The same way DOS was replaced by Linux for people that did not need Windows, AJAX/browsers will be replaced by Flash / Apollo / shells
As far as ease of development – I learned my lesson about it as well. Simplicity has its price – moving thinking process toward the end of the project. Understanding the application has more impact then the tools you have at hand. Applications that match AJAX toolkit are rare – and will become even more so. The times of AJAX are now, which means that pioneers already left the space 2 years ago.
I also believe that the best tools for most of the tasks are the ones you know best or made yourself – just keep learning. It also helps to re-read Joel Spolsky essays or “Mystical man years” from time to time to put things in perspective