Under My Palapa

I saw some people were finding this blog by doing a search for ExtJs with JSF.

I previously searched and found that someone had wrapped up ExtJs into a JSF library, however it appeared to be a dead project.

A coworker and I instead decided to use Apache Shale to expose methods in a managed bean, to return the JSON data that ExtJs wanted. At the time, we were only using the ExtJs Tree component. We needed to use the ExtJs Tree instead of the one from RichFaces because of a certain requirement that could not be fulfilled with RichFaces.

Looking at the beginnings of a new webapp. I’m heavily leaning towards Grails+ExtJs, even though the ExtJs plugin has been discontinued because of the license change. Honestly, paying for developer licenses is not a big deal in the grand scheme of things. For widgets, dojo seems to be a runner-up, however it just doesn’t feel as polished, and I find the experience somewhat lacking. I wish jquery had a more mature & official widget set, because it would be my first choice.

In my previous project, we used JSF 1.2 + RichFaces. While I feel that combination can produce good results, it didn’t feel very agile. On a new JSF project going forward, I would recommend Seam; however it is not worth the effort to retrofit a large project.

James Lorenzen has a post about how to roll your own ExtJs plugin for grails, which I will probably give a shot.

With JSF, we were pushing markup down to the browser using ajax polling & rerendering of components. However, I want to start using RESTful services along with Comet. This should get interesting…

My evaluation of JavaFX is fairly positive, I think it is a step in the right direction. I do however have questions about Sun’s commitment in this area. I get the feeling their developers are understaffed and overworked. I am currently setting JavaFX aside and will take another look at it closer to its actual release, either beta or release candidate status. I could definitely see using it in a small, self-contained project. However right now I can’t recommend that it be used in an enterprise/commercial app in the foreseeable future. It comes down to tooling — the visual tools need to produce a much better looking result out of the box.

The entire developer/designer workflow thing is something of a question mark. My belief is that most organizations are currently not really employing designers to begin with, so getting into this mode of development is going to be difficult. This is why making it easier for engineers to produce something that looks nice should be the 1st priority IMO.