Blogging Roller

Alexis Moussine-Pouchkine links to XULFaces, a JSF renderkit for XUL. So, can you really develop a webapp, switch renderkits and have it look like those screenshots? Gotta admit, that's pretty cool.

There have been a number of somewhat confusing announcements and threads regarding the open source licenses that are popular among Java developers: APL, LGPL, MPL and CDDL. Here's my take on a couple of those items.

Open office goes LGPL

The first announcement regards LGPL and SISSL. Sun has retired the SISSL and the Sun-sponsored Open Office project will be making all subsequent releases under the LGPL license. This doesn't mean that Sun is moving en masse to LGPL as at least one Java blogger seems to think. It just means that the OpenOffice folks decided that LGPL is the best license for their product.

JSF goes CDDL

The next announcement regards CDDL and JSF. Sun has released its Java ServerFaces implementation under the CDDL. Some Java sites covered this as news, but I don't think it is all that newsworthy. Sun had already released its J2EE implementation Glassfish under CDDL and J2EE now includes JSF, so most license savvy folks probably already assumed that JSF was under CDDL.

Apache and the LGPL

And finally, we come to the touchy topic of Apache and the LGPL. Apache is close to deciding a policy to define how LGPL components may be used in official Apache releases. Does this mean that Apache projects can now use and ship LGPL components. No, not yet. Cliff Schmidt announced the <a href= "http://mail-archives.apache.org/mod_mbox/www-legal-discuss/200508.mbox/%3c96D75844-B15A-45E3-BF8D-5CF4517A09E2@apache.org%3e">proposed policy and that policy was to be voted on by the Apache board on Aug. 17, but I've heard no news of this.

In case you're wondering what this might mean to Roller, the proposed policy allowed Apache projects to depend on third-party LGPL components but not to ship them. For Roller, that would mean that we can keep on using Hibernate, but we'd have to ask people to download the Hibernate jars separately from Roller. That's unpleasant, but it's not too bad, I guess. If you read the email thread that Cliff's email started you'll see that some folks would like to allow Apache projects to ship third-party LGPL components, but not to have hard dependencies on them. For Roller, that's actually worse. We'd have to support an alternative non-Hibernate implementation of the Roller backend. We've been there and done that (i.e. supported both a Hibernate and Castor/JDO backend) and it was not an experience that I'd like to repeat. I wonder what happened in that Aug. 17 board meeting.

I'd also hoped to answer the question posed by Geert Bevin on The Server Side: What's the best developer-friendly open-source license for Java products, but that'll have to wait for a later post.

Rafe Colburn: I don't know about you, but I'm much more like to port my PHP applications to Ruby on Rails than my Java applications.

To me, Ruby is a lot like C#. Not in terms of syntax or language features, but as a pleasant improvement over what was there before. See, if I ever have to do any Windows programming again, I'll be thankful for C# and .Net. They're a great improvement over C++/MFC and Visual Basic. And likewise, if I ever have to work in the world of the scripty web P languages, I'll be thankful there's Ruby. It looks like an improvement as well. And like C#, Ruby looks pretty easy to learn.

Raible's published his first magazine article (contrats Matt), an insightful and informative piece on the state the J2EE web tier for JDJ titled Challenges in the J2EE Web Tier While Frameworks Driving Innovation. He traces the motivations and history behind JavaServer Faces (JSF) and the many challenges it faces.

I'll admit, I hadn't really given much thought to the mismatch between JSP and JSF. After reading Hans Bergsten's excellent article Improving JSF by Dumping JSP again, I tend to agree with Matt. A cuddly little open source project called Facelets could be the savior of JSF. And check out the Facelet docs; they've been getting some help from JSF spec lead Ed Burns. When I get back to my RollerFaces experiment, I'm going to give Facelets a try.

Raible: If you're hosting your open source project at java.net, you might be annoyed by the fact that you can't disable the forums or issue tracker. The forums pretty much suck, mainly because there's no "remember me" on the site and it's a pain to click 3 times just to answer someone's question. They also don't have the ability to quote a previous message. In addition to the less-than-adequate forums, the issue tracker doesn't hold a candle to JIRA.

For the longest time, I've been wanting to remove both the "Discussion forums" and "Project issues" links from the AppFuse homepage. I asked the java.net folks, and they didn't have a solution. Finally, I thought of one on my own.

Matt offers some constructive critcism of java.net (and by extension CollabNet).

One of the other things I did over the break was to start reading Core JSF again. I've made it about halfway through this time and I must say, the book is quite good. I like the way Geary and Horstmann walk you through the learning process. Starting with something small, simple and not IDE specific and slowly adding in the pieces you need to build a real app. For me, that approach dispelled the "JSF is too complex argument." I'm finding that JSF is a lot less complex than Struts, which is really all I have to compare it against. I might actually finish the book this time ;-)

At JavaOne, I picked up a copy of Java Creator 2 early access (EA) and I've been playing with that as well. I was hoping to build a simple JSF front-end for Roller, just for fun. Creator looks great and the form designer works well, but when I noticed how many com.sun.* classes end up in the generated code I backed off. According to the <a href= "http://developers.sun.com/prodtech/javatools/jscreator/ea/jsc2/reference/docs/Installation_ReleaseNotes-ea.html"> release notes the whole JSF implementation and com-sun-web-ui-appbase jars are redistributable (as are mail.jar and activation.jar -- that's news to me), but the idea of IDE specific code makes me uncomfortable -- especially in an open source context. So instead of using the form editor, I decided to base my little RollerFaces experiment on example code from Core JSF, which so far relies only on the javax.faces packages.

I decided not to use the form editor, but I didn't want to give up on Creator completely. So I tried to use Creator in place of Netbeans, but ran into another problem: Creator 2 EA only supports the Sun Java App Server and the built-in Netbeans Tomcat launch/debug feature has been disabled. That's disappointing. I don't want or need a full-blown EJB app server for this simple learning excercise. I want Tomcat. It's light-weight, starts fast and I know it well (as do most Java webapp developers, I expect). According to the Creator EA 2 docs, Tomcat support is coming soon. Personally, I would have supported Tomcat and simple webapps first and then Sun Java App Server and more advanced EJB stuff later. Please the most folks first, right? Anyhow, I backed off of Creator 2 EA entirely and now I'm hacking RollerFaces in Netbeans 4.1. I'll try Creator 2 again when it hits beta.

My second JavaOne was a great experience, but it was a little stressful because up until last night I couldn't find any of my co-speakers. I spent most of Wednesday preparing to give the whole talk by myself, but luckily for me (and the attendees), Pat and Kevin showed up just in time. Unfortunately, Pat showed up with some very bad news for us at Sun: he's leaving to work at Google.

In the end, I think the talk went pretty well. Kevin did most of Pat and my slides on syndication because we had split the talk 50-50 when we couldn't locate Pat on Wednesday night (and assumed he was still in Paris). He did a good job with the material and added in some interesting points from his experience at Rojo.com where they parse millions of feeds per hour with the Java-based Apache Commons (sandbox) FeedParser.

We were a little disappointed with the turnout. I'd be surprised if the 700+ seat Yerba Buena theater was more than 30% full. The fact that were in a lunchtime timeslot on the last day of the show certainly didn't help. Anyhow, I'm relieved that it's over and ready for a nice long week off.

Are there really going to be any smack-downs? I'm betting it's gonna be pretty mild. I'll blog the comments that provoke the smack-down bell. Webwork - Jason Carerra

• "We're not desgined by a committee"
• "You have to join the JCP and press the spec meat"
• "You have to have a tool to deal with the 18 JSF configuration files"
• "JSF, which hasn't been out for a year, is telling us about maintainability?"
• "Yeah, use the JSF extensibility to create a spyware component"
• "(Geary's mailbox is so full because) people are confused by MyFaces"

Struts Shale - David Geary

• "JSF is a lot better Struts and a better Webwork"
• "Struts is old school, get over it" (he asked for the bell, but didn't get it)

Wicket - Martin Dashorst

• "No framework makes it as easy to create components as Wicket"
• "With Wicket you only need to know the extends keyword"

Tapestry - Howard Lewis-Ship

• "It's not Struts++"

JSF - Ed Burns

• (too nice to provoke the bell even once)
• "You don't care about maintainability, use Struts" (shoulda gotten the bell, but didn't)

Audience

• To David Geary: "Why did you steal Springs web flow?"

OK, I missed a couple, but you get the idea. Those guys were waaaay too nice.

Lots of cool news during the general session this morning. I took lots of notes, but I've got no time to type them all in. So, here are the high-lights from my point of view (and in the order they were mentioned):

• Java in every DVD player. The new Blue Ray standard for DVDs (up to 50GB storage per disk) includes Java technology. I don't have the details, but word is that all Blue Ray DVD players will include a Java VM for creating interactive content, i.e. the menus, special effects, and games that accompany movies on DVD. Gosling mentioned that future DVD players will have network ports too. Sounds like lots of interesting opportunities for Java developers there.
• IBM supporting Solaris 10. The entire IBM middleware suite will be available for Solaris 10 (and I'm assuming OpenSolaris as well) in both AMD and Intel flavors. This is great for Eclipse-heads that dig Solaris (not that I would know any of them ;-)
• Sun Java App Server now open source and under CDDL, a true OSI certified open source license (with patent protection/indemnification). Jim Driscol has some more details of this new project, known as Glassfish.
• Screamin' Sun box with ALL dev tools is $29.95/month. I'm tempted to get one myself. You get Sun Studio C/C++ dev environment, Java Studio, and Java Creator with tech support for less than thirty bucks a month -- all preloaded on one of the new Sun Ultra 20 AMD Opteron based workstations.
• Java Creator 2.0 EA and AJAX components available: Tor took the stage with Johnny L to demo Creator 2.0 and a cool new AJAX component they've cooked up (I'm still trying to find it on Java.Net).

Pardon (and correct) me if I don't have all the details right. I don't have time to write up Graham Hamilton's presentation on Java SE, EE and ME. Maybe later. Gotta run...

Simon Phipp's and Jim Grisanzio have organized a <a href= "http://blogs.sun.com/roller/page/webmink/20050621#free_beer_at_javaone">JaveOne bloggers meetup at the <a href= "http://maps.google.com/maps?q=661+Howard+Street,+San+Francisco&spn=0.035205,0.059566&hl=en">Thirsty Bear from 6-8pm on Monday June 27th. See you there.

I'm attending JavaOne 2005 next week and speaking on Thursday at noon. The talk is titled Beyond Blogging: Feed Syndication and Publishing With Java™ Technology (<a href= "https://www28.cplan.com/javaone05_93_1/session_details.jsp?isid=270318&ilocation_id=93-1&ilanguage=english">TS-7318) and there are four speakers Alejandro Abdelnur (not attending), Pat Chanezon, Kevin Burton and myself.

Preview version of the new NetBeans CVS Support now available!: The NetBeans team is proud to announce a preview (alpha) version of the new redesigned CVS support, an innovative approach to integration of a Version Control System (VCS) into the NetBeans Integrated Development Environment (IDE). The new module provides a simple, efficient IDE versioning workflow and intelligent support for common versioning operations.

The module is not derived from the current NetBeans VCS support, but instead is meant as a lightweight approach with less focus on architectural abstraction and much more attention to User Interface (UI) functionality.

CVS integration is the one thing that keeps pulling me back to Eclipse, so I'm really glad to hear the VCS subsystem is being re-worked (maybe 'replaced' is more accurate). Check out the screenshots, looks like good stuff. I hope better Subversion support is not far behind.

Via Roman Strobl and John Jullion-Ceccarelli

Wouldn't it be awesome if you could just apt-get, pkg-get or emerge Java^TM on your favorite open source *nix platform? Wouldn't it be a wonderful thing for Java developers if Apache HTTPD came with built-in support for J2SE and J2EE? I think we just took a step in that direction.

Tim Bray outlines SMEX-D, a really simple way to describe by REST and SOAP based web services. I like it. I'd like to see the SMEX-D for Atom Protocol right about now. Compare to Norm Walsh's NSDL.

Java 2 Platform Standard Edition (J2SE) 5.0 Release 1 for Tiger

I want it, but when can we ditch JDK 1.4 support in Roller?

Weiqi Gao posted an informative and insightful article about the recent GCC 4.0.0 release, which features a much improved version of the GCJ Java compiler and GNU classpath, a free as in FSF implementation of the Java class libraries. I don't agree with his predictions, but he's done his homework and put together bunch interesting links.

Joseph B. Ottinger: So far... "Wow" and "Well, it's okay."

I found the installation to be a breeze, which isn't too much of a surprise. It found the various JVMs on my machine, including 1.5, which was good, I suppose.

It looks absolutely smashing. (That's good.) The development pane is easy to understand at first glance, also good (although note that I'm not a rank newbie to Netbeans, so my prior experience may factor in here.) It's pretty fast, too, which is good to see. Maybe this will help shut up the SWT hordes.

Lately, I've been working on making the switch to Netbeans 4.X (I'm using both 4.0 and 4.1-beta). My expectations were pretty low due to my past experience with Netbeans 3.X, but Netbeans 4.X is great. It's a whole new IDE, the Ant integration is awesome, and 4.1 adds the refactoring I need to make the switch. Plus, it's fast. I have to agree with Charles Ditzel, the tables have turned and now both Netbeans (and IDEA) seem to be faster and more responsive than Eclipse.

Unfortunately, my Eclipse addiction is still not entirely under control. The one area that needs work in both IDEA and Netbeans is the source-code control integration, I still have to fire up Eclipse when I need to synchronize with the CVS repository. I feel like I'm flying blind without Team Synchronization view.

I just noticed, DataDirect (based here in Raleigh) has blogs -- ok, just one, but it's a start.

Patrick Peak's been off writing <a href= "http://www.amazon.com/exec/obidos/ASIN/1932394419/theartofwar-20/102-4163380-5716141">Hibernate Quickly for Manning, but now he's back with a slick new Roller theme. He's written a nice couple of posts attempting to debunk some of the Ruby hysteria that's been infecting Java bloggers of late.

We took Jetbrains up on their offer of free IDEA IntelliJ IDE licenses to open source projects. Today the licenses arrived and I mailed them out to the Roller committers. I'm happy to have another IDE to work with because neither Eclipse nor Netbeans completely satisfies me on Solaris x86. Swing-based Netbeans and IntelliJ run perfectly on Solaris x86, but unfortuantely Eclipse does not.