Posted by editor
on January 19, 2006 at 7:19 AM PST
Catching bogus input with WebWork... also:
Feature Article: WebWork Validation
Projects and Communities: Location-sensitive drag-and-drop and Weblets
Also in Java Today: SOA best practices and dancing with the event-dispatch thread
Forum postings: Numerical computing and is UDDI still alive?
Weblogs: Architecting Swing apps, the content-type header, and clarifying "dependency hell"
Catching bogus input with WebWork
WebWork is a web
application framework designed to keep productivity high and the
code simple. It has gained popularity for several reasons, including
its integration with Spring , a powerful tag library, and OGNL
support. Its powerful validation framework is borrowed from another
OpenSymphony project, XWork .
WebWork doesn't get as much attention as some of the other webapp frameworks, and that's too bad, because its adherents think it gets a lot of things right. Its most recent release clears things up for its upcoming merger with Struts, where it will presumably get on more developers' radar.
In today's Feature Article ,
WebWork Validation , Zarar Siddiqi looks at how to do web-form validation with WebWork, setting aside simplistic approaches like hard-coding validation inside the
execute() method in favor of reusing WebWork's built-in validators and writing your own custom validators. He also briefly notes WebWork's client-side validation capabilities, which are provided by another java.net project, Direct Web Remoting .
In Projects and
Shannon Hickey's blog entry Location-Sensitive Drag and Drop in Mustang illustrates changes made to drag-and-drop in Mustang, allowing potential drops over a JTree to be indicated by showing the drop location only when the mouse is over a valid location. "Prior to Mustang, developers could not implement this very important behavior due to oversights in the Swing implementation."
From the Java Enterprise Community : "The Weblets project announces the availability of its Version 0.1 release, and has graduated out of the Java Enterprise community incubator. Weblets makes resource file management and versioning as easy for Web development (for example, for JavaServer Faces component library developers) as it already is today for desktop-based Java development. application frameworks."
In Also in
Java Today ,
the interview SOA Best Practices: A Conversation with Sun Microsystems Distinguished Engineer Mark Hapner discusses the goals and realities of building service-oriented architectures today, covering standards (and why not to wait for them), the importance of XML, the relationship of SOA to web services, and advice for SOA programmers.
Dealing with Swing's event dispatch thread is a challenge for some Swing developers, but Scott Delap still thinks you should Learn to Dance with the EDT (When Debugging Swing) . Responding to comments compiled by Alexander Potochkin's Blog , which call for automating the transition of calls to and from the event dispatch thread, Scott writes "Desktop application development is this very choreographed dance. On one side you have the predictability and simplicity of running all UI logic (painting, input handling, etc) on one thread (the EDT in the case of Swing, the display thread in the case of SWT). Then you mix in the complexity of having worker threads for long running tasks and the asynchronous complexity that this brings. Automating this dance is not impossible but not trivial either. Your code ends up only as good as your defenses."
In today's Forums ,
ylzhao seeks solutions for
Numerical compuing in Java
Recently, I want to write an application, which involves some numerical computations like matrix, linear algebra and linear and nonlinear optimization, equation root find etc. Traditionly, numerical computation is done by Fortran or C language and some commerical softwares. However, if I use Java language, then I need to use JNI to communicate with the library files. If not, I should convert the librares written in Fortran or C language to Java, which is not so easy. So my question and request are: Does JDK team or JCP group consider to add some standard numerical computation packages to JDK in the future?