The Swing Application Framework (JSR 296) aims to do for client-side Java applications what Web frameworks have done for server-side code: Abstract out common application patterns with the goal of making development easier and less error-prone. In this interview with Artima, Swing project lead Hans Muller talks about this new JSR.
http://www.artima.com/lejava/articles/swingframework.html

In this interview, Hans Muller shares what he thinks are the roughest edges of Swing for developers building rich-client Java applications. Do you share his views of the problem, and the solutions proposed for JSR 296? What do you think are currently the hardest aspects of rich-client Java development?

Looks like Sun is transforming the JCP into an SCP, the Sun-community-process. From now on other companies have nothing to say. Welcome to dictature!

All those JSR's are Sun-only without prior consulting of other projects! Eclipse is build on top of the OSGI-framework. So JSR 277 is part of Sun's cold-war-strategy against the Eclipse-project. But JSR 277 is only the tip of the iceberg. There are more JSR's like that: JSR 295 (data-binding) and JSR 296 (Swing-application-framework aka „Netbeans“ included in the JRE). All those JSR's are „Sun-only“ and Sun didn't consult any project or developer outside of Sun itself or outside of the entirely Sun-controlled java.net-projects.

I agree in part with Arthuro, in that I've heard little about existing deployed application frameworks (Eclipse RCP, Spring RCP, NetBeans Platform) when this new JSR has been brought up. A question for Hans would be--while you have commented on the NetBeans Platform as regards this new JSR in the past, I've heard nothing about the other two (they appeared in one slide from your JavaOne talk). How are they influencing your design, discussion, approaches? What have you learned from them and what have you rejected?

My main concern is whether this will be yet another JSR without any public review of ongoing discussion, where the expert committee emerges from seclusion at some date and drops a specification on the front porch, then disappears. This behavior is typical of all the JSRs I've seen and I find the whole "community" marketing angle a little cynical. These are JSRs developed by "experts" for the rest of us. I don't count myself as an expert in any arena, but I'd much prefer an open discussion board/mailing list where at least we could follow the progress of the development. Politicians got used to sunshine laws, there's no reason programmers can't as well.

And yes, I know there is now a JSR to "reform" the JCP. Of course, who knows when that will bear fruit or what fruit it will bear.

I'd rather see this JSR developed under an open-source umbrella project as with SwingLabs. The SwingX project community is generally both open and hard-nosed about design, and the end result is generally better because of it.

That said, I like what Hans has to say and am glad he's leading it. I like how he presents his ideas and how he shares opinions when he does blog or comment on something.

Regards
Patrick

I've hesitated to respond to this claptrap since it's obviously a
troll. Unfortunately there's always the possibility that someone else
who hasn't read the interview will be mislead by it. So, just for the
record: JSR-296 is not an attempt to wedge any existing application
framework into Java SE. Not the NetBeans (rich client) Platform, and
not Eclipse RCP. The JSR-296 expert group includes developers who are
intimately familiar with both of these IDE-based frameworks and I'm
sure their experience will help us do a great job defining our own,
relatively narrowly scoped, desktop application framework.

[This is a response to Patrick's post. Apparently the artima.com
forum software doesn't splice replies below the post that you're
replying to!]

The reason I only brought up the relationship between NetBeans
Platform and JSR-296 is just that's what I was asked about. Although
I'm not interested in playing Dr. Frankenstein and stitching together
parts selected from worthy existing application frameworks, we are
looking at other people have done. All three of of the frameworks
you've mentioned are much broader in scope than JSR-296 and most of
what I said about the difference between JSR-296 and NBP, applies
to them too.

> Looks like Sun is transforming the JCP into an SCP, the
> Sun-community-process. From now on other companies have
> nothing to say. Welcome to dictature!

Can you substantiate this claim?

In the interview, there is a link to the JSR's home page:

http://www.jcp.org/en/jsr/detail?id=296

You can easily see that JSR 296 has 11 expert group members, in addition Hans, who is the spec lead. So out of 12 people, to my knowledge, only one represents Sun. In fact, you will see JetBrains on the EG as well, which is a Sun competitor in the IDE space.

When posting to these forums, please make sure to check your facts first, because it saves others time, and also because you can make more constructive comments that way.

> I agree in part with Arthuro, in that I've heard little
> about existing deployed application frameworks (Eclipse
> RCP, Spring RCP, NetBeans Platform) when this new JSR has
> been brought up.

As Hans has already mentioned, the interview used the NetBeans platform as an example for a desktop application framework. Because of the scope of the interview, we did not have time to compare JSR 296 to other platforms.

I'm not a troll. But I am concerned about Sun's arrogant behaviour (spell NIH) towards the Java-community. Why don't you consult other projects before you start your own JSR's? Are you aware that i.e. the Eclipse-project is working to get databinding done for both SWT and Swing? Why don't you consult those projects?

I want to see Hans Muller, Scott Violett and the likes discussing with other developers and projects before you throw your Sun-only JSR's upon the whole Java-community. If this behaviour continues I see the Java-community being splitted.

So why doesn't Sun's Scott Violet consult the Eclipse-team doing a Beansbinding/Databinding-project right now? Read that transcript with Scott Violett and Hans Muller: http://java.sun.com/developer/community/askxprt/2006/jl1016.html

That's odd. Scott Violet equalizes Swing with Netbeans if you know what I mean. Read exactly what he says. The whole interview is nothing but a Netbeans-marketing-plug disguised as Swing-talk

There are more people being concerned about the way the JCP is governed by Sun Microsystems.

Read the comments on JSR 277:
http://weblogs.java.net/blog/stanleyh/archive/2006/10/jsr277_early_dr_1.html

Hans: thanks for your reply. I was posting mainly because when this JSR has been brought up, the existing frameworks I mentioned (apart from NB Platform) are not mentioned in the same breath--I'm glad to find out this is a misunderstanding on my part and that the expert group is taking feedback from people familiar with other frameworks.

Also, I want to say on a positive note that this JSR is very timely--I'm looking forward to the time when I can roll new Swing apps quickly without having to decide on all the (many) issues involved in setting the application up, and without having to hand over the keys to the App Overlord.

In my work, Swing has been a target a couple of times, but they were always internal projects so I couldn't take any framework code with me, and I don't have the extra time to develop such a library on my own. So being able to get new Swing apps up and going quickly will be a relief.

Some questions:

1) I'm wondering if, apart from the framework's RI, will the expert group be releasing some small, real-world, demo applications? That would be welcome--perhaps as a companion FOSS project where community members can participate.

2) I believe the JSR doesn't cover a plugin architecture; is that correct?

3) Will there be an "upward" migration path to enable applications to scale up to use some of the larger platforms if necessary?

4) Do you feel the databinding problem should be taken as a separate concern, or are you keeping the binding JSR in mind when developing the application framework?

Keep up the good work, I apologize if my tone was too negative in my first post.

Cheers
Patrick

Is there a web page where we can find more information about JSR 296, download specs, tutorials, examples, JARs etc?

I'm particularly interested in retargetting applications for different platforms (Windows, Mac OS X, Linux etc.) whilst maintaining compatibility with the platforms' own conventions. Does JSR 296 address this?

Steve

One of the biggest frustrations in swing is that listeners
need to be unregistered before components that the listeners
are registered to are garbage collected. How does JSR 296 address this ?

One good reference for Swing is

http://www.java-swing-tutorial.com

Salman Khan
<a href=http://www.google.com/>Salman Khan</a>
http://www.google.com/
www.google.com