RSS 2.0 New Entries Syndication Feed Atom 0.3 New Entries Syndication Feed

Show blog menu v



Use it



Sibling projects

RIFE powered

Valid XHTML 1.0 Transitional

Valid CSS!

Blogs : Archives

< FishEye rocks   Prominent Java developers praise the RIFE framework >
What's next for Rich Internet Applications?

Face it, the scope of web applications is expanding. Text-based information sites will still continue to use regular XHTML, but a lot of new application types are appearing now that Rich Internet Applications (RIA) are a reality.

Personally, I'm standing behind the open-source OpenLaszlo platform. The main reason is that consumers are adopting computers and the internet for most of their multi-media needs. Nobody finds it weird anymore that you listen to music or watch a movie on your computer. Portable music players like the iPod go hand-in-hand with desktop applications to manage them, applications for home-recorded movie editing are included in most operating systems and Apple's Garageband brings awesome audio creation power to everyone. OpenLaszlo allows you to write RIA that incorporate sound, music, images, video and text easily, just look at what Pandora has done.

The main problem with RIA is their runtime platform. Flash is the most appropriate technology for it nowadays due to its market penetration and singular target. Sadly, there are major speed and stability problems. Anything beyond simple applications need careful architecting and exhaustive testing to run acceptably. Alternatively, Ajax may seem nice for certain effects, but it will always remain an uphill battle to track down all the browser differences and work around all the specific platform bugs. I thought we learned that lesson with DHTML six years ago. Also, Ajax also totally lacks any kind of multi-media capabilities.

I'm actually disappointed that the proliferation and integration of Java applets has failed. I'd trade any Flash solution for Java2D and Swing. If only the download size of the JRE would be smaller and the installation as transparent as Flash. Sadly I think that Java's reputation in this area has totally been compromised and that any user runs away as soon as the word applet is dropped. While Java Webstart brings a good alternative for full-blown applications for a targeted audience, it doesn't integrate transparently with the browser. Maybe Sun just has to repackage applets and give it another name, touting it as the next big thing and doing it right this time.

The recent canvas feature of Safari and Firefox looks promising. They seem mostly compatible and there's a standard in the works that specifies how they should behave. Maybe it could be used to build consistent RIA user-interfaces which use standard Ecmascript for the logic and only use Flash for the multi-media fragments. Now Internet Exlorer only needs to play along ... :(

I wonder what's next for cross-browser and platform agnostic Rich Internet Applications. What are your predictions?


I added two movies, one with Wingz (which uses applets) and one with OpenLaszlo (which uses Flash). Make up your own mind which one is the most user-friendly. Note that I use this on MacOSX where Java is installed and integrated with the browser by default.

The movies require Quicktime and TSCC Codec, which you can download for free.

QuickTimeWingz version (3.2MB)

QuickTimeOpenLaszlo version (2.2MB)

posted by Geert Bevin in Laszlo on Sep 15, 2005 11:33 AM : 29 comments [permalink]


Re: What's next for Rich Internet Applications?
Great minds think alike! I think Server Side JavaScript applications will be making a come back especially with the likes of Mozilla Rhino and the new Mozilla browsers support for E4X. There is also growing interest in using Direct Web Remoting in AJAX applications.

I have only just finished blogging my thoughts on a similar theme, Desktop Applications using JavaScript engines and The Return of Server Side JavaScript (SSJS).

As for an Applet replacement, have you seen Pluglets? I'm not sure they'll take off but they are certainly an interesting idea.
Re: What's next for Rich Internet Applications?
mm..i really think you listened in on the conversation I had on IRC where we were considering that OpenLaszlo might make an output to something like canvas and/or svg
Re: What's next for Rich Internet Applications?
You see that I listen to you when you say something ;)
Re: What's next for Rich Internet Applications?
Pluglets indeed look interesting, but here the cross-browser problem pops up again. It's not viable for a RIA to only target Mozilla browsers.
Re: What's next for Rich Internet Applications?
i tried to answer on tss but cannot get through somehow
<blockquote> Alternatively, Ajax may seem nice for certain effects, but it will always remain an uphill battle to track down all the browser differences and work around all the specific platform bugs. I thought we learned that lesson with DHTML six years ago. Also, Ajax also totally lacks any kind of multi-media capabilities.</blockquote>

I dont think you understand what ajax is about. The only difference is that you update certain parts of the page instead of the whole page. What you are actually saying is that html lacks mm capabilities. Wich is IMO not the case.

Laszlo is nice to look at but a pain to develop with. If they come with something more dynamic i will give it another try.

To get back to ajax again, there seems to be a shift going on from serverside programming to client-side programming in correlation to ajax. This is not good!!! We should still code on the server but being able to script the client ON the server. There is huge difference between these two. All ajax should do for us is a little automated dhtml in a generic way so that we can update certain parts of the page with serverside commands. The whole thing should not take more than 20 lines of javascript and should be totally hidden from the developer. And we should drop markup, or at least have it as an add on not as "the only way". I cannot say it enough, at first you seem to be able to deliver quickly but you bump your head eventually when times goes by. Why is eclipse so damn stable, fast and maintainable? Because they don't use markup. We should generate markup from objects, not build classes around markup. If you work like this your whole view technolgy is pluggable.

If you manually type your markup your stuck with it forever, the bigger the application gets the further you move away from being able to ever replace it with something else. I mean we interface everything we can think of but not our view technology , why is that? Maybe a bit off-topic but i find it an intersting subject. I also don't like translations, i prefer implementations. Implementations are faster more clear and maintainable.
Re: What's next for Rich Internet Applications?
Hi Dennis,

I thought about specifying that Ajax distinction in my post, but then didn't. Of course, Ajax != DHTML, but apparently when people talk about it they always include the more advanced DHTML features. This is logical, since merely updating parts of the DOM asynchronously is not was makes an application a Rich Internet Application. Dynamic popups, drag and drop, animations, fades, sliding, .... all those things are totally normal and only belong on the client-side.

HTML does lack multimedia capabilities without plugins. There's no way to have animations, play a movie and interact with it or do the same with music by only relying on HTML without having to install anything else.

About Laszlo, I initially had the same apprehensions as you, just read through my blog ( As time went by I was pleasantly surprised by it though, as you can read further on.
Could you please explain what you mean by 'something more dynamic'?
Re: What's next for Rich Internet Applications?
Geert (are you dutch?),

I did not mean to post my previous thing twice , something went wrong there;)

I agree upon the fact that html does not support mm out of the box, but is flash not a plugin? I really dont care that windows media player is a plugin, or quick time. It is easy to interface them, no big deal. To tell you the truth about sliding and moving stuff, after a while it is just anoying. It degrades performance of the application. People that work with applications want the application to work as fast as possible, they are not waiting for sliding and moving stuff. The css property display(block/none) is a wonderfull thing, it lets you fold and unfold anything you like. I build truly rich stuff with my own framework with only a few lines of javascript and i can choose between
- FO (wich again gives me awt,pdf,text,postscript etc)
- HTML , laszlo (to slow though)
- XUL(mozilla and SWT)
- WML.

Writing another view technology implementation cost me a few days and all the work i have done so far catches up automagically

By laszlo not being dynamic i mean that the pages have to be compiled (jsp like) after they are changed. Just because of that only i cannot work with it, as i cannot work with jsp. I wrote a laszlo interpreter for laszlo that lets me work in a dynamic way but that is just way to slow for serious stuff. This doesn't mean that laszlo is crap, it is one off the nicer things out there but it is not enough for me for now.
Re: What's next for Rich Internet Applications?
I'm Belgian ;)

Sure you can integrate any plugin (which people have to download separately), but Flash is almost as widespread as any browser. The thing is that it remains a plugin if you just use it to play the animations or the movies, there's no new interface paradigm for the web application that allows you to interact with the movies, photos, animations, or sounds directly. The GUI widgets are also relatively restricted.

Don't get me started on XUL, that held great promise for us and we we seriously burned by it. We started an application in it and ditched it to rewrite everything in Swing with Webstart.

Did you try the latest version of Laszlo? They dramatically improved the compilation speed and reloads are now almost instant. It's damn close to being interpreted.

Do you have more information about your own framework, is it available somewhere?
Re: What's next for Rich Internet Applications?

some people say that client/server architecture will be back for business multimedia applications, with the help of citrix and others...
and that WWW will stay like this for a while...

I'm not sure they are right...

But the real problem is "standardization"...

Micro$oft is the only one to have 90% of the browser over internet...
So.... we must move with them and create a nice standard
with mozilla, apple... but if they don't want to do something...

IE 7! will have tabs. So, could imagine improvements like this
in ie 16 ?

But maybe the solution is a mix between flash ajax and applet...
A Fla-ax-plet ? :p

What do you think about this ?

Re: What's next for Rich Internet Applications?
I've written two articles about this subject and how I think web development is going down the drain and won't help people as much as other solutions. The articles are here.
Re: What's next for Rich Internet Applications?

Thanks for the XUL warning i' keep it in mind. I think your story proves my point that manually typing markup is a really really bad thing.

I will give laszlo another try tonight, looking forward to it;)

My framework is available on a sneak preview basis for the ones that are really interested. It will be fully available when it is 99% stable and 100% documented. I can work like a mad scientist at lightspeed with it but that will not be the case for people that don't know how to work with it. If you are intersted i give you an account and you can ask me anything you like to know about it and i'll be happy to answer all your questions. One thing i can say about it is that it opens java files in eclipse for you when you click on a widget in the browser. No more endless puzzling, code is always one click away.

let me know if you are intersted.

Met vriendelijk groet;)
Re: What's next for Rich Internet Applications?
I'm certainly interested in your framework!
Wingz vs laszlo
For me java/swing and laszlo could not be compared.And each one af it's one advantage.

With swing or wingz you could provide true application, with native look and feel. Lot of Enterprise application do not require flashy look and toy interface.

For sum-up it's not the same market.Wingz or swing , for me is for B2B and flash or laszlo are perfect for B2C.

For be clear, for our service department we already done some project with lazlo.But for Business inteligency dashboard or Enterprise application our client prefer wingz or swing application.

Re: What's next for Rich Internet Applications?
Flash was never intended as a GUI, it's an animation library. You notice this as soon as Flex is thrown against it and performance goes down the drain.

AJAX is nice, but it still seems a battle against all the browser differences. But it can go a long way, because it is so widely used.

Java works like a charm. I did an applet with Hessian as the communication protocol to the back end (more Java). The front end only contains the display logic and effort was put into making it look good with gradients and rotating windows (yes all Swing). Indeed the JRE download is large, but RIA focuses on higher bandwidth anyhow and at the rate broadband is increasing, this is not a concern. The Java plugin makes sure applets have the same caching functions as webstart. The demo is a total succes aside from one problem: normally the applet is up and running within 10 seconds, but on very incidental machines it can take up to an hour! And I have no idea why.

Re: What's next for Rich Internet Applications?
Does anyone here have an opinion of macromedia flex? It has quite a spiffy api, but my concern would be reliability and speed for it.
Re: What's next for Rich Internet Applications?

I like to send you a link to the framework i mentioned earlier.

Re: What's next for Rich Internet Applications?
Hi Dennis,

you can mail me at gbevin[remove] at uwyn dot com.


Re: What's next for Rich Internet Applications?
Wow, these are very brave notes about the future of RIA… :'(

Here are two of somehow related aspects should be considered:

* Future of the MyIDE or the best for developers and users … and
* Future of the technology that will be supported by the industry leaders.

I do not discuss now the second aspect (too much politics and money talk would be involved), but regarding the first statement… You probably should not discuss the future of the RIA without mentioning of IAB Studio

IAB Studio is a RIA server with IDE, reporting, data management, charting, workflow and various business modules. This product makes the development process and usage of the applications easy, comfortable, confident and provide end-user with exceptional experience.

IAB Studio is available for FREE download with documentation and many examples.

Additional information, sample pages, demo application and other resources are available at
Re: What's next for Rich Internet Applications?
IAB Studio looks like one of the worst things I've ever seen. No Safari support, and in Firefox the layouts are very flaky, your default theme looks horrible (I hope there are others), and your site contains no valuable information at all.

Actually, I showed your screenshots to some people before, to illustrate how badly GUIs can be designed, as an example of what not to do. Are your developers color blind?

And ... is your whitepaper really just a just one HTML page with a tiny little blurp of text full of buzzwords?
Re: What's next for Rich Internet Applications?
Hello Geert,

I'm wondering if you managed to start it all up.

Re: What's next for Rich Internet Applications?
Hi Dennis,

thanks for the mail but I haven't got time yet to try it out. Everything is a bit hectic with JIA in less than 2 weeks and EuroOSCON in a month. I'll try to make time for it, but I don't know when, I still have a lot to do. It might be during my flight to Florida. Is it tested on MacOSX?
Re: What's next for Rich Internet Applications?

It is developed on osx. I had to get used to it but cant go back anymore now. You probably know what i mean. Take your time. I tried rife/laszlo but nothing seems to happen after logon, this is also the case with the web demo. Rife/crud is kind of the same idea i orginally started with, i was so tired of doing the same stuff over and over again. And besides that there was always something small wrong, spend half my time (if not more) on assembling my application instead of coding it. Anyway, does days are gone as more and more people seem to see the light(ness). Using jetty instead of tomcat also helps as you also discovered;)
Re: What's next for Rich Internet Applications?
Hi Dennis,

RIFE/Laszlo isn't supposed to *do* much in itself. The purpose is the same as RIFE/Jumpstart: get started building a new application in a matter of seconds. Everything is set up for development and runs out of the box. The functionalities are thus deliberately small.
Re: What's next for Rich Internet Applications?
Hi Geert, would be curious to know more about this:

"Don't get me started on XUL, that held great promise for us and we we seriously burned by it. We started an application in it and ditched it to rewrite everything in Swing with Webstart."

Re: What's next for Rich Internet Applications?
We were building an application with a lot of forms with it. Performance dropped so badly that even on my Powermac Dual G5 2.5GHz it was unusable. In Swing you barely notice it updates.

There were also some technical problems that made handling the XML with the population data very tedious, but I have to ask about detail for that if you want to know more.
Re: What's next for Rich Internet Applications?
I like the way you think about desktop java... ;)
Is a 5MB JavaPlugin really too much to download?
Re: What's next for Rich Internet Applications?
I'm a fan of Laszlo.

A huge win is the built in ability to obtain (and send) XML to / from a Web Server (REST services are becoming a best-practice here) and to bind the results *directly* to UI components.

No parsing of the XML and creating or populating tables / trees / inputs with the results, you only need to bind the XML "dataset" (or part of it using a limited XPath syntax) to a component, and the rest is automagic. Changes in the XML on your page, caused perhaps by user interaction, is reflected automatically by the bound components... deleting nodes causes views to disappear, adding or moving XML nodes around in the DOM rearranges things on the screen.. it's beautiful and easy.

Check out the components available

And pay particular attention to the Databinding sections
(laszlo in 10 minutes/data/data binding) although, you don't get a full appreciation of databinding through those examples.

You can compile the Laszlo code into a standalone Flash application which needs none of the Laszlo components for deployment. My Laszlo apps perform very well, integrates seamlessly with the rest of the app (using Struts / Tomcat) and the line between Flash and DHTML is not at all visible to the users...

Once you've tried Laszlo and become accustomed to how it works, you won't go back to AJAX / DHTML based complexity. I literally can do complex UI's magnitudes faster than I can without it.

My best-practice is do the easy stuff with DHTML (Struts, JSF, Maverick, whatever) - but to do the complex & interactive sections (with lots of data) using Laszlo... it works.

Re: What's next for Rich Internet Applications?
Have u guys tried running Laszlo applications against the new Flash Player (version 8)?

Flash Player 8.5 looks interesting with its new "ActionScript Virtual Machine" too
Re: What's next for Rich Internet Applications?
No, we can't get version 8 to install on MacOSX Tiger, so we can't try it out. Version 8.5 definitely looks very interesting.

Add a new comment

Comments on this blog entry have been closed.

< FishEye rocks   Prominent Java developers praise the RIFE framework >
Google web