[jimk1723]

Heh. Dot Not indeed. There's plenty of comparisons on the interwebs. That said, Telenic's Rad (they're rad!) components didn't really knock my socks off.

If you're looking for shiny, you might want to check out OpenLaszlo or Adobe Air (The RichFaces dudes are working on integration with Flex as well. Flamingo, I think it's called)

I wouldn't be wooed to another development environment by the promises of an $800 set of UI components, it seems fool-hardy at best.

[sergeysmirnov.sim.exadel.com]

Formal correction: RichFaces dudes do not have any relation with Flamingo project. Exadel, Inc. does. Non-JSF solutions are not play fields for RichFaces team.

What we really did (I did, actually) was investigation of integration between Ext-JS and JSF (in terms of creating JSF components based on them).

Some kind of result:
http://livedemo.exadel.com/richfaces-rex-demo/

Conclusion: (Short version) Ext-JS is buggy and hardly integrated with JSF paradigm of development.

[krkempa]

First of all: that was an angry post written after I hit problems with RF once again. It was a little bit too harsh.
I see the value you are bringing to JSF, Nick fighting with bugs in JSF and FF internals. Please continue, but improve the way you are doing it.

Talking about development objections... Let me explain (concretes 2 acapites below, but please read all). Two years ago we started with Tomahawk components. Soon after, we realized that most of these components are unfinished and hardly work. Instead of stabilizing these code they were pushing new features that were sharing the sins of those actually implemented. These objections were shared by other developers (Seam developers amongst them). I brought the idea of using JSF to my dev team but after first whoaaa!, people started to become dissapointed by the quality of components for our pilot effort.

Then Richfaces got opened and I was very happy because I knew Exadel from many educational materials on JSF and perceived this company (as well as Icefaces guys) as only possible saviors of the JSF components market.

Now I'm worried I see the same thing that was with Tomahawk happening to RF. Please understand that we (the Seam and RF communities) try to build live sites using your components as a base. We need this base to be as solid
as possible because often our visitors can judge us after some malfunction of our site. And we need to do version upgrades, because RF is maturing and we need those fixes up there...

I was happy to see Christian implementing this wiki in public but please remember that users of this site are of a quite unique type. We are mostly developers who knows Seam and RF, so we understand why/that something is not working. This is not the case with other types of sites.

Talking about development once again I don't know what to say if day after one of the 3.1.x GA releases I see errors after looking at you demo site for only 2 seconds, brought by this release. Dissapearing attributes and numerous regressions between versions introduces very discrete bugs to our platforms.

Please, instead of adding new blink stabilize what you already have. Modal panel needs love (or a total rewrite rather) - there are repeatable bugs questioning its performance and the way it kills MacOS' browsers. There are other bugs in basic components needed for every RFied site which can kill developer even before he get to shiny calendars, effects etc. Please be more pragmatic when introducing new componets. Value of sortable table which can sort Integers only after toString() values is rather limited but instead you can clean and fix the code of other components.

Please, improve your regression tests suite, make more complex demo with interactions between components. Maybe the problem with vanishing attributes in your tags is somehow related to the way you generate(?) your components using CDK? Maybe this is the place you can place additional tests?

Roleplay a situation when someone upgrades RF on a site already visited by many users. Things like populated to caches unversioned resources and ResourceNotFound exceptions can kill a site soon after an update. This exceptions still happens in 3.1.x branch as I see on JIRA...

Check what TELERIK has (that's the correct name, I made a mistake before), there is a big hype around on what their Javascript can do and how good it is. And good how it works on a HUGE pages with hundreds of components. I saw it deployed on a site my friend is currently deveoping. It makes me wonder where we are hading...

Check the performance and results of using you components on such extreme, nontrivial situations. You demo site is not enough...

Thats my (big) 2 cents. Hope things will change.

Thanks,
Kris.

[krkempa]

I'm glad you have narrowed down the problem :)

[chris.simons]

Uh just to add my two cents to this debate (which I have just found.  I admit I am a little late...

We switched from IceFaces to RichFaces due to some compatibility issues and bugs with Seam.  We also figured that moving to RichFaces (another JBoss product) was the right choice to ensure future stability and compatibility.  I can't say that I'm overly disappointed with RichFaces but I have had many issues with components rendering completely wrong in major browsers (IE 6/7, Firefox 2/3).  It seems that their releases have focused on adding new components which in turn may break existing ones - it is definitely true and they have taken a lot of heat for these issues.  I think the RF team is working resolve them as we speak and - from what I've read - is focusing on resolving these types of issues before introducing any new components.  It is a shame really because we've married RichFaces to our application in such a way that to replace it with something else would be a 3 month task in of itself.

I think RF will come around; Telerik can deliver amazing components for .NET because they are mad expensive.  The RF team doesn't have the luxury of licensing components out at $1000 a pop.

[jnusaira]

We like richfaces but as others point out ... welcome to whats broken today

... i mean come on it took a while for them to fix the i used a calendar object on my page and it causes my safari browser to crash ... that was inexcusable

And many of their components you have to use hacks if you want them to appear dynamically via Ajax.

Thanks to the OP ... i will have to check out those that you mentioned. 

[jamesjmp]

I also chose RF because of RH, although I had a look at IceFaces as well.
I must admit ExtJs demo is fascinating but in my project switching would mean too much work, and I guess that it happens to more developers, some of them have already stated that.
I think RF have improved and I think they will come around with stability. I´m really looking forward to that after facing so many problems.
About the future, who knows.. right now I think components should come together with JSF, and regarding that I´m not sure if ExtJS would be the best solution. But as developers we all do know that we always will have to keep an eye on emerging technologies that eventually would devour their predecessor...

[chris.simons]

Agreed - rich:calendar is too slick for its own good.  We've had lots of problems with this component in particular.

[taccart.thierry.accart.name]

Hi Kris

to answer your question 'what do you say',
I personnaly answer that people including classes of javax.servlet.\* in a jar they named gwt-user.jar (delivered in gwt-windows-1.4.62) are [put here some bad words] ...
As if they would not care about java architecture, standard, and rest-of-world :-(

Maybe am I not objective, because I personnally loose days of investigation on a deployment problem ^{(I'm ashamed to write it!)} : I did not thought one minute that jars from GoogleWebToolkit, supposed to be deployed on application servers, could contain javax.servlet.* !!!

Moreover, I rely on Hibernate, on JBoss, and I'm sure that problems coming from youngness of RichFaces will be fixed in short term. Of course, with JBoss I don't expect the same service level as with BEA Weblogic, but the price is not the same.

My final words will just raise a question : how many of us are using open source projects ? how many have submited ideas, reported bugs, or participated to bug fixing or even to development ? :-)
 

[marx3]

I'm too disappointed by instability of RF. It took half a year to release stable version with properly scaled columns of rich:datatable (while one of snaphot had it). We have to use snapshot in production environment...
I don't see ext-js as an option. It seems more buggy than RF...