Wednesday

Wednesday began with a morning meeting, followed by the 11:00am "JavaServer Faces Component Vendor's Panel". The panel consisted of representatives from ILOG (JViews Charts, Diagrammer, and Gantt), Otrix (WebTree, WebMenu, and WebGrid), ESRI (ArcGIS Server), Software FX (Chart FX), and Business Objects (Crystal Reports). Most of these components were demonstrated inside of Java Studio Creator, but one vendor's components were shown from within Eclipse, and another didn't demo their components inside an IDE at all.

I was pretty familiar with most of these products, with the exception of Crystal Report's JSF support. Their suite of components make it quite easy to log in, navigate through a repository, and select, view, or display reports using convenient JSF components. It's nice to see JSF support added to Crystal Report's stack. Craig McClanahan ended the session with a demonstration of additional components announced in Java Studio Creator 2 - specifically a new grid component.

Next up was "Shale: The Next Struts", presented by David Geary and Craig McClanahan. If you're not familiar with Shale, you can think of it as a bag of goodies for building JSF applications. It includes lifecycle events for backing beans, page flow support, command chaining, mock objects for testing, and a new Tapestry-style templating system called "Clay". I'm pretty familiar with Shale, but I hadn't seen any examples of Clay until this presentation. It's nice to see some innovation happening around JSF's display technology; you can find Geary and McClanahan's sample application at Geary's Core JavaServer Faces web site.

For some strange reason, the pavilion floor always closes the day before the end of JavaOne, so Wednesday was my last day to check out new products and meet some vendors. As usual, I made the best of my time, visiting Exadel (Exadel Studio 3.0), M7 (NitroX), Oracle (JDeveloper and ADF Faces Components), Simplica (ECruiser DataTable, runtime, and other components), ILOG, Quest (JClass ServerChart), ESRI, and others. The two primary Eclipse plug-in JSF IDE vendors -- M7 and Exadel – both had great offerings.

I have spent some time working with M7's NitroX for JSF previously, and it does a great job of melding a powerful WYSIWYG designer with a quality web development tool. The killer feature of NitroX, however, is its AppXRay technology, which has a keen understanding of how all of the resources within your web application (configuration files, JSPs, and Java code) are related. So if you ever reference a class that doesn't exist in faces-config.xml, or use a non-existent managed bean in a JSF EL expression, NitroX will let you know.

Exadel's next release of Exadel Studio also has great visual support for JSF, including an embedded version of Gecko (Mozilla's HTML rendering engine), and the innovative ability to update the view as you type each character. Nice!

I also spent some time talking to two rich-client vendors – Isomorphic and InsiTech. Isomorphic sells SmartClient Web Presentation Layer, which is an Ajax-based component and application framework. What's cool is that you can define your UIs via XML that lives on the server, so you get the deployment benefits of the web with a very rich user experience. (The UI is updated dynamically, so the page isn't constantly refreshed). Isomorphic's product is well suited for JSF integration, which I mentioned that to some of the people at the booth - hopefully they took it to heart.

InsiTech's XTT takes a radically different approach – all of their applications are pure Java and use Swing. You can still define the UI via XML, but an applet is responsible for downloading the definition and rendering the view on the fly. In contrast to the Ajax-style approach the result is truly rich client interfaces.

I took a short break from my pavilion tour to grab some food and see a glimpse of the Web Framework Smackdown, which consisted of a jovial group of fellows: Ed Burns (JSF), David Geary (Struts Shale), Howard Lewis Ship (Tapestry), Jason Carreira (WebWork), and Eelco Hillenius representing a new framework called Wicket. I didn't see too much of the session, so I can't tell you who won, but TheServerSide has a nice overview in their JavaOne coverage (which includes details about other JSF sessions too), and the users of Javalobby had something to say as well.

Next, it was time to attend the MyFaces gathering at the Thirsty Bear. It was no surprise that Oracle was sponsoring this dinner, since they officially announced that they are going to be contributing their ADF Faces Components to the MyFaces project, and joining as committers. It was great to see all of the MyFaces players I had seen earlier, as well as some others, such as Martin Cooper (of Struts fame) and Mark Raible (of Simplica). I have worked with Mark before, but I had never met Martin; we had a good chat about music, which is always a fun topic for computer geeks.

I left the MyFaces dinner in time to catch most of a 10:30pm BOF about JSR 273, The Design-Time API for JavaBeans. This JSR's goal is to "extend the JavaBeans specification and APIs to improve design-time functionality for use within IDEs." This is important for any type of component that's used inside an IDE, most notably Swing and JSF components. The current JavaBeans API isn't quite up-to-date when it comes to specifying how components work when a developer is using them inside an IDE's visual designer, so most vendors end up extending the APIs in a proprietary manner. The JSR uses the work done for Sun Java Studio Creator's JSF support as a starting point.

This JSR, in conjunction with JSR 276 (Design-Time Metadata for JavaServer Faces Components) should greatly reduce the amount of effort necessary to integrate components into different IDEs. I'm a member of the JSR 273 expert group, but it hasn't begun in earnest, so I was anxious to hear what Joe Nuxoll (the spec lead) and his colleagues in the Sun Java Studio Creator team had to say. Afterwards, I joined some members of the Sun crew for some late night grub at Mel's Diner.

Page:   1   2  3   4