I was a big fan of Flex. The code was clean, object-oriented, re-usable. We even had the bonus of E4X. We’ve wrote the interface of a huge project using it and I know there was no way we could have had such a rich client other than using native code. We developed multiplatform and the client ran multiplatform without a bit of change. The deployment was easy and the administration on the client side minimal. Everyone had flash.

When Adobe dropped the bomb it struck us from head to toes. That’s what happens when you depend on a proprietary platform, someone said to me. This was some time ago and I hopped things will advance in the mean time. Yet Adobe is still going the bad way and in the mean time I see the less and less RIA choices. (2 paranthesis here. First I know Adobe is still saying that it’s commited to having flash on windows but who can still depend on it? and second when I say RIA imagine something really complex like an accounting system not the average iPhone app, there is still a need for that kind of corporate applications).

JavaFX? I like java but who the hell would choose an Oracle product nobody seems to use which Oracle can shelf faster than Adobe did considering its much smaller userbase and the love of Oracle for non-profitable products.

HTML5? The wonderful kid, the hope of us all is 20% tech and 80% hype. Yes, try your bakery app but would you try a product which is supposed to take 1-2 years to develop and be in production/maintenance/support for 5-10 years afterward? There are a few projects which promise to be THE solution but they could as easy go the way of the dodo bird. And in the mean time you see all the html5 failures (1, 2), you see browsers changing automatically each month or so. And I understand why building something on this base is like playing cards with the devil, there is no way to win no matter the effort. And yes, the effort would be enormous no matter which wonder library will use.

Swing? Given the above a better choice might seem the ugly and even more-hated-than-flash java web start solution using an OpenJDK alternative. Can a solution from 10 years ago, which never got much wide spreading be the best we have now?

Native? Or even better, go native, go broke writing and maintaining different clients. Say goodbye to web loaded content, easy deployment and write-once-run-anywhere code. And more, hope not in a few years to depend on proprietary, walled market to deploy your application.

Have we gone backward or, more probably, I am missing a big size of the picture here?

Comments:

Zsolt Plavetz -

IMHO, GWT (Google Web Toolkit) is the solution.


len -

:( I was facing the same choice 5 years ago: GWT versus Flex. I choose flex and for some time I was very happy. But the Spark components lack a lot compared to the MX ones, especially in mobile and I have not seen any big move on GWT part. Bump for me :(


bogdan -

http://www.infoworld.com/d/application-development/the-long-death-of-fat-clients-196554


len -

Sunt de acord ca sunt multe chestii pentru care JQuery este de ajuns. Dar exista totusi o nisa de applicatii pentru care ce poti face, in momentul de fata, nu e de ajuns. Plus ca este inca o iluzie sa crezi ca un cod scris acum va merge peste 2-5 ani. Ar merge un cod scris acum 3 ani in JQueryUI pe toate browserele?


Zsolt Plavetz -

… combined with Adobe AIR.