Blogging Roller

Alejandro Abdelnur: My belongings are in storage, they've been there since March, got a ticket for next week, a small backpack, a camera, a diving computer and about 4 months to spend in Southeast Asia.

It's sad for Sun to lose such a talented and motivated developer (Alejandro is one of the guys behind ROME, among other things), but it sure sounds like Alejandro has got his priorities in order. Good luck and safe travels!

I updated the Roller User Guide for Roller 2.0 today. The new user interface and group blogging features are documented with plenty of screenshots. Comments, corrections and suggestions for improvement are welcome.

In case you live under a rock and haven't heard the news yet, Netbeans 5.0 beta is out. I'm downloading it now. I tried one of the q-builds a couple of weeks ago and was amazed by the new CVS integration. This time around I'm going to take the new Matisse GUI builder for a spin and see what it can do for my clunky little BlogClient.

Quite an interesting interchange today between James Robertson, Scoble and others about the state of blog tech standards. Like many developers, James thinks the legacy specs RSS, MetaWeblog API, and OPML all suck and he's not afraid to say so. Legacy defender Scoble shoots back that he's a user and therefore he doesn't care about specs. Nevermind the fact that Scoble started the conversation by demanding that tools implement a specific spec, OPML. Anyhow, Winer joins in and says, it's a user vs. developer thing and Scoble set that mean developer man straight.

I think Scoble and Winer are right, it's about the users. When you create a data format or netwok protocol specification, your users are the developers who have to implement the spec. In the case of blog tech specs, the users think the specs suck. That's why the many incompatible variants of RSS are being replaced by Atom format and MetaWeblog API is being replaced by Atom protocol.

• JBoss and Microsoft
  That's the "big" announcement?
• Web 2.0 meme map
  Apparently the so-called Web 2.0 is just a rorschach test, what do you see?
• Not economic integration?
  "the latest laughable Times report about those 'schools that work'"
• Raible evaluates CMS options, Pt. I
  Raible evaluates Java, Perl and PHP CMS options
• Raible evaluates CMS options, Pt. II
  He notes that none have blogging features as good as Roller
• JDS and OpenSolaris
  "goals for the desktop community, why OpenSolaris was important, and what our plans were for JDS."

After getting some feedback from Alejandro Abdelur of the ROME project, I spent part of my weekend revising my Atom 1.0 patch. I sent my new patch to Alejandro and to the ROME dev list today.

• What is TailRank?
  Kevin Burton: "TailRank is my new startup"
• Business Journal on Converge South
  "conference that will bring prominent journalists and bloggers to Greensboro "
• Economic integration
  Raleigh effort "most ambitious in the country to create economically diverse public schools"

• Not So Simple
  Interesting article on RSS ads, mentions Sun

RIFE, a continuations-based web application framework for Java has decided to use the CDDL license. I like the way Geert puts it, CDDL: a copy-left license without ambiguities.

Geert Bevin: For the 1.1 release of RIFE, we have decided to dual license the framework and add CDDL as an option alongside the previous LGPL license.

The CDDL offers a copy-left open-source license that doesn't have any ambiguities regarding the terms 'linking' and 'executable', which make no sense in a Java application. There is no possible viral behavior that extends beyond RIFE itself. The CDDL protects us by requiring all source modifications to RIFE to be contributed back under the same license, but imposes no restrictions at all on the use of RIFE in a commercial application.

However, to be able to combine RIFE with a GPL license, a stronger copy-left license is required. In this case, the LGPL can be selected.

Cool. A blog about Raleigh. Via Wade Minter

I don't want to brag of course, but it appears that Roller has the most complete Atom Protocol 04 implementation around. As Tim Bray notes, my server was the only one on the air for yesterday's Atom interopathon. We tested post, put and delete of entries and my implementation was a little brittle in places, but it held up. Roller's Atom implementation is in the sandbox and won't be included in Roller 2.0 because the protocol is not yet done. I'm hoping for Protocol 05 real soon now.

Why do I care so much about Atom Protocol? Simple. I can't finish my book without it.

Here's the mildly historic moment that Tim mentioned:

Since I posted yesterday, I've learned that Subversion is in the works for Java.net. If your new project can't wait for that, you can try Javalobby's new JavaForge service.

Matt wonders when Java.net will provide Subversion based source code control.

My post never got through to the ROME dev list, so I posted the patch file, a Groovy example and the patched ROME jar to the Roller wiki: Rome08Atom10Patch.

In preparation for the Atom Protocol interopathon this week, I've been busy updating my Atom Protocol client and server implementations. They were using Atom 0.3 because they're based on ROME and that's all that ROME supported, but this weekend I modified ROME to support Atom 1.0. I submitted a patch to the ROME team, but for some reason it still hasn't shown up on the ROME dev mailing list.

I wrote to the Google Blog Search team and they say that the service *does* index Atom 1.0 feeds. No word on when they'll start to deliver seach results in Atom 1.0 format instead of the deprecated Atom 0.3 sillyness they're serving up now.

I just signed up for both days of Converge South 2005 over in Greensboro. Looks like conference, which focuses on "journalism and multimedia web blogging," will be well attended.

Welcome to Elias, who brings the Roller committer count up to seven (in chronological order: Dave, Matt, Lance, Anil, Henri, Allen, and now Elias). Elias runs one of the biggest Roller installs in the world at IBM and he's got lots of cool feature ideas and practical suggestions for improving Roller.

Mailing-list traffic and interest in Roller really seems to have picked up recently. I'm not sure if that's due to interest in Roller 2.0 or the ongoing move to Apache, but either way that's a good thing for Roller developers and users alike.

On this day one year ago I was raving about my first week at Sun. If I had the energy to write a long retrospective blog entry, I'd brag about Will Snow's amazing team, all the Roller revs we've shipped, speaking at JavaOne, winning the chairman's award, and Roller's ongoing move to Apache, but I don't. I'm completely drained by a day of documentation work and two hours of flash-cards with my 7-year-old. So I'll just say this: I love working at Sun.

Now that Google is indexing newsfeeds and providing Google Blog Search, is there any hope for the wee little blog search companies like Technorati, Feedster, and PubSub? They'll be OK, don't you think? After all, nobody would have invested in them if they didn't have a contingency plan for Google's entry into the market.

It's interesting to me that the servce returns results in "funky" RSS 2.0 format (RSS 2.0 with Dublin Core elements for date and creator info) and Atom 0.3. I assume they can index any format, but I wonder: are Atom 1.0 feeds indexed by Google Blog Search and when will they start producing search results in Atom 1.0 format?