What to do about Java and JSF? 
December 12th, 2006
As an Ajax component vendor we support multiple backends (ASP, ASP.Net, PHP, Cold Fusion, JSP) but we’ve been wanting to support our Java/J2EE customers better for a while now. We’ve done a lot ad hoc research, and it seems clear JSF is not the way to go. According to some, like Martin Percival from BEA that’s selling a Eclipse based JSF IDE, it’s not the way to go:
One thing that never fails to surprise me in the developer community in Europe, is the spectacular lack of folks actually using JSF. Struts is still riding high in terms of implementation and knowledge. Is it just that JSF is too hard? Talking after the event, people were suggesting that they thought it’s use is more prevalent in the rest of Europe but this does not tally with any of my ad-hoc polls carried out at various events this year. I can almost guarantee a 1% show of hands for JSF usage and no more.
Any suggestions? Start with basic POJOs and Servlets, then Struts? Are we stuck in some kind framework buzz frenzy…and no one’s using any of it? We’d like to know:S
Technorati Tags: ajax, jsf, java, pojo, servlet, bea, ide, martinpercival, j2ee
Del.icio.usThis entry was posted on Tuesday, December 12th, 2006 at 9:10 pm and is filed under Technology, Software Development, AJAX. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.
Entries (RSS)
December 15th, 2006 at 7:12 pm
I can tell you that we use JSF. I doubt that is relevant to decision making, but to echo your point JSF is not simple. However, none of the web frameworks are. So, what I have found is that I have an easier time getting junior developers up on JSF via tools (formerly Creator now Netbeans Visual Web Pack) than I could trying to teach them a framework. In most cases, I have to write components, small libraries, or give them recipes (to avoid having ten different solutions to discovering which rows in a table are checked). However, it does seem to be a better approach overall and still gives me the flexibility to dig into the framework to do fancier things when needed.
I haven’t bought your product yet and I am still debating it (though I probably will). However, I think that it would be a no brainer if it worked with the design time jsf api. Not sure if that helps or not, but keep up the good work.
December 17th, 2006 at 6:34 pm
Thank Les,
I really appreciate you taking the time to comment on this. So I take it you’re using Netbeans then…were you using Creator before it became Visual Web Pack?
Have you looked at the JMaki project that Greg Murray is working on? It promises to allow you to use Ajax components within Netbeans…and could developers like you and I a lot of time when building components for others. Let me know your thoughts.
It does seem like a no brainer for the design time api in jsf, it just seems there’s a number of different IDEs and frameworks, they’re all slightly different and mostly tricky to work with. It’s more a question of which to support first…
Thanks for reading and I hope you become a Nitobi customer! If there’s anything I can do to help let me know.
Andre.
December 17th, 2006 at 11:42 pm
We have just started on a development project using JSF with Sun Studio Creator.
The framework does have a relatively steep learning curve, but the features of Sun Studio Creator (or Netbeans with Visual Web Pack) allow for extremely rapid development.
We are very interested in the Nitobi Grid, and (like Les mentioned above) hope to see the release of a version with design time support in either Creator or Netbeans.
Keep up the good work.
Ian
December 20th, 2006 at 12:12 am
Hi Ian,
Thanks for the feedback! Is it the visual IDE tools or the JSF framework that provides the most value to your dev process? We’re trying to get enough customers together to move forward on this! Let me know anything else that would help you out.
Andre.
December 22nd, 2006 at 10:44 am
Thanx for the response, Andre
Yes I was using creator prior the the visual web pack. In fact, I suffered through using some of the very early releases. As soon as it came out I jumped on it simply because I was tired of the complaints about webwork/hibernate being too difficult (which, I thought was easy…i think they had xml-fear-itous). :) We are on the visual web pack now, which is a very welcome release. We were having to run both creator and netbeans side by side which was a bit memory intensive.
I’ve looked into jmaki and have waffled back and forth on it. It seems to be a good idea on the surface, but they have not built adapters for some of the more complex components yet. So, I have passed so far based on whether or not it can handle things like dnd or complex components like trees and tables.
If you guys are struggling with ide support, JSR 273 is designed to take care of that problem. From the looks of it, there have not been any updates in a while, but the spec lead and submitter are two of the javaposse guy. They have indicated in their podcasts that it is still moving forward. Maybe you might want to drop them an email and see if they have a preview of that api.
When I get back to some ui work (probably late january…have a bunch of backend integration work to redo right now), I will be picking up your components. I think that the vispack design time api is pretty straight forward. So, I may take a shot at wrapping the combo box, at least. The table my be a little harder.
I am interested in knowing, though, about your thoughts on other components that you are doing. I have an interest in a good tree impl with dnd. So far, dojo is the only one that supports dnd insert lines. I have been following jack slocum’s work over at yahoo and have been very impressed with the extensions he has done on his blog. You may want to consider integrating some of his work (not sure about license). Either way, keep up the good work.
LES
January 3rd, 2007 at 10:24 pm
Andre, you asked Les:
> Is it the visual IDE tools or the JSF framework that provides the most value to your dev process?
For my group it’s definitely the visual IDE tools. I’ve been looking *everywhere* trying to find a good set of HTML/RIA/AJAX tools that support visual design. It’s a requirement I have for whatever I end up choosing. The few tools I’ve found have other significant problems–e.g., their demos run slow as heck, and if the demos run slowly….
If you were to support visual design tools, it would be a good differentiator for your products. (And your demos are not slow as heck, which is also a huge plus.)
My team was originally going to use Java Studio Creator but is now looking at NetBeans.
January 4th, 2007 at 8:14 pm
Hi Brian,
Thanks for the comment. I’m stoked you think our demos aren’t slow as heck! I would think the new visual web pack in NetBeans would be the logical choice over Java Studio Creator as it’s pretty much the evolution of it. Why are you considering this over Eclipse? I’m guessing it’s primarily the integrated visual tooling but I could be wrong.
Have you looked at Jmaki running in Netbeans?
Ping me offline if you’d like to chat more. Tool support is on the block for 2Q and 3Q this year.
Thanks!
Andre.
January 4th, 2007 at 8:40 pm
@ Les:
All these frameworks, components and IDEs have various levels of ease of use. What we truly need in the Java is a slick integrated solution that’s as easy as building VB apps in Visual Studio. Otherwise we’re cutting out a huge segment of developers who aren’t going get anywhere.
We’ve looked at JSR 273, we’ve even done some beta level integration with JSC last summer. It’s for an older version of Grid now but you can get it here: http://www.nitobi.com/jscdownload/. I’d be more than happy to share the code with you so you can try to wrap it. I think we have some code for our Ajax combo box also. Email me and I’ll get it to you.
Finally we do have several new components coming out in our suite this Feb. An Ajax tree is one of them. Check out our roadmap plans here: http://www.nitobi.com/about/roadmap/.
Thanks!
February 15th, 2007 at 3:15 pm
We looked at JSF but the generated HTML was too crufty and the usability guys just hated it. So we decided to go with Struts 2 for our new project. It looks pretty good so far but we are running into a few configuration issues. Hopefully that is because it is still maturing.
Scott