Blogging Roller

Pat Chanezon responds to James Snell's proposal for collaboration between ROME, which is (in my mind) the premier feeds API for Java, and Abdera, a newly proposed Apache project that aims to implement the Atom format and protocol.

Seema explains how to get the latest Roller code running on Sun Java System Web Server 7.0 Preview 1. And Seema's right, Roller 2.3 (incubating) should be out real soon now. RC4 is looking good and I think it might be the gold.

• Jonathan Schwartz: Java Jedi?
  "I would really love to see Jonathan and the people trying to change Sun succeed"
• Sun stuns server market in Q1 with sales spike
  "Sun regained its position as the big-iron leader"
• MetaWeblog API for MSN Spaces
  Via Dave Winer: MSN Spaces now supports MetaWeblog API
• Sun Portal is now Open Source
  "Liferay, eXo, JBoss time to watch Sun driving open source Portals market"

SAS Institute is still shying away from blogging, but Sun's Phillip Kinslow recently started a blog about his job, which is to align Sun with the SAS mantra "The power to know." So we've got SAS employees blogging about Eclipse and Sun employees blogging about SAS; how about some SAS employees blogging about SAS? There are lots of cool things going on at SAS and I bet the employees are just dying to tell us about it, but they need a push. Plus, they need a policy, some servers and some software.

PS. Sun's BM Seer is also blogging about SAS and some world record setting ETL benchmarks by SAS Enterprise Data Integration Server and a Sun Fire E25K.

Congrats to continuous integration guru Jared Richardson, who has jumped ship from SAS and apparently without much of a safety net. I don't know Jared personally, but I heard a lot of good things about him when I was at SAS -- so good luck to Jared.

• Apache Geronimo is still troubled
  Ken Coar is concerned that Apache Geronimo dev not completely open
• An Attempt At a REST Programming Model
  "an easy way to expose resources from POJOs and interfaces" -- via REST
• ongoing · The End of SOA
  "SOA isn’t the future, Web style is"
• Will Java EE 5 bring REST to Java?
  "On one hand the support for REST within Java EE is good news..."
• Redeemed! JAX-WS still sucks!
  "I have no illusions that JAX-WS will be scrapped in favor of something completely different"
• Does Java EE 5 getting REST mean WOA will break out?
  "[with REST] Interoperability is automatic and guaranteed" -- does anybody really believe that?

After I got back from JavaOne, I loaded up on all of the cool new software I saw in action at the conference, including:

• Netbeans 5.5 beta (build 200605090801)
  
• Netbeans 5.5 Enterprise Pack
  
• Netbeans 5.5 Subversion Plugin (installed via Update Center)
  
• Java Studio Creator 2 (Update 1)
  
• Java EE 5 SDK: includes Sun Java App Server (SJAS) 9 plus Derby and other goodies

I installed all of this stuff and I've been using Netbeans 5.5 with Subversion support all week. On my Solaris/x64 box, Netbeans 5.5 seems very stable. The Subversion client is a little flakey, as is to be expected for pre-release software, but it's been really holding its own during some refactoring and package renaming work I've had to do this week.

So, if you're itching for Subversion support in Netbeans, give the 5.5 beta a try and help the Netbeans guys out providing feedback on the mailing lists and issue tracker.

The JavaOne general session demos last week (e.g. the build-a-blog-server in 5 minute demo) seemed to show an extremely fast build-deploy-test loop. So next, I'm going to try to switch to SJAS9 + Derby for development and debugging and see how it compares to working with Tomcat / MySQL. Last time I tried SJAS (version 8) for development, I found it to be a little too heavy for my tastes.

Sam Ruby, James Snell and Robert Yates have proposed a new Apache project to provide a reference implementation for Atom: the Atom Reference Implementation or ARI. The plan is to implement an Atom parser, client and server libraries.

Code is already available from James Snell's site (ari.tar.gz, under IBM copyright) and, currently, there are two simple examples: one for parsing a feed and one for posting an entry. I've already taken a quick look and the code looks very good. But it does require Java 5 generics, so we couldn't use it in Roller right now now even if we wanted to.

ARI will definitely overlap with ROME to some extent, because ROME provides an Atom format parser and has plans to include an Atom protocol client and server toolkit. But ROME has always focused on supporting all feed formats (e.g. RSS 0.9X, 1.0, 2.0, Atom 1.0 and even the old and deprecated Atom 0.3) and that's not the intent of the ARI.

Some little notes from JavaOne in roughly chronological order:

• Netbeans Subversion plugin looks really good (and I'm using it now)
  
• Netbeans Maven integration is deep and wide
• Those MacBooks can be pretty flakey (and my sympathies to Jason)
  
• For open source Java, the question is how, not if
  
• You can now apt-get Java on Debian, Ubuntu and other Linuxes
• Sun Portal, Creator and JMS to be open sourced very soon via Java.Net
• The Java Petstore has been updated to use AJAX and Java EE 5
  
• Sun is helping with Eclipse for Solaris/X86
• Disney is using Roller for internal blogs
• The Geronimo guys like to use Roller to demo ease-of-installation
• The Java Web Server guys like to do the very same thing
• Google Web Toolkit compiles Java code to AJAXy JavaScript UIs
  
• Tim Bray's Sigrid is a sweet little Java grid engine (but where's the source?)
  
• The JMaki project brings AJAX (Dojo, Scriptalicious, etc.) to JSP and JSF
• Visual Basic for Java is on the way via Sun's Project Semplice
• Slides for all JavaOne talks are now available online, audio coming soon
• Webcasts of general sessions are available too
  

Danese Cooper: Earlier this week we all heard about Jonathan Schwartz and Rich Green hinting they were about ready to release Java under some OSI-approved license. Supposedly they just need to nail down "How to Deal with Compatibility". I read this news with some irony, since I know that they bloody well know exactly what to do already. Its been discussed every year since 1999 inside of Sun. Their covenant with Apache and the Geronimo has already successfully demonstrated that it can be done (compatible FOSS reimplementations of Sun-generated specifications). They are simply being disingenuous. What they really mean is "How can we placate the FOSS community without giving up control?" which is the age-old question for Sun.

Give us Sun-folk a little credit. If you want open source Java runtimes (and I do), then the announcements this week were most definitely good news. Up until now, we didn't have our story straight. Jonathan Schwartz was telling people that all Sun software would be open source and the OpenSolaris folks were showing us how it could be done, but Java leaders like Gosling (and many others) seemed to be saying that Sun's implementation can never be made open source for fear of incompatible forks. Now we're all on the same wavelength. That is a good thing.

And I don't think anybody really knows "exactly what to do." I'm more than happy that the Java community is going to figure out how to do it right. I believe (and I think Danese would agree) the JCP needs some work and more participation before it can get us all working in harmony (no pun intended) on compatibility.

Tags: Open Source, Java, Harmony, JCP

(Usual disclaimer applies: I do not speak for Sun)

(photo thanks to Marc Hadley)

I haven't seen any reviews yet, but I thought my talk went quite well. Over 500 people registered and it seemed that the room (capacity 620) was overflowing. My demo worked, despite the fact that I had to work on a Windows box (and found some problems in my .BAT scripts). And I got the timing right too. I finished about ten minutes early, but that was perfect because there were lots of questions.

I think I got the point across that Atom protocol is generic, not just for blogs and applicable to a wide range of problems. I also made a point of promoting Marc Hadley's work on WADL and his talk on RESTful web services with JAX-WS.  Turns out Marc Hadley was in the audience. We'd never met before and my talk ended at noon, so we had lunch together and had a nice chat about WS-* and REST and life in general.

I attended Marc's talk later in the day and it was packed; I'd estimate he had over a thousand attendees. Based on the attendance at my talk, Marc's talk and the number of times I've heard or seen the acronym REST mentioned, I'd have to say that Java folks are pretty interested in putting the web back into web services (is that a Jon Udellism?).

You can get my slides from the JavaOne Content Catalog (link is TS-1756). You can also get the software that I demonstrated, which was the Blogapps Server (the Atom server) and the Chapter 10 code from the Blogapps Examples (the Atom client) download. You can get both of those from the Blogapps project on Java.Net.

Tags: Java, JavaOne, REST, Atom

I'm here in San Francisco and ready for JavaOne. Time for some more horn tooting: one more reminder to come see my talk on Tuesday morning. Here's the info:

• Sam Ruby: Blogging with Style
  Looking for "a simple set of checks which could be made to enable style attributes to pass safely through feeds"
• Feed Manager for Movable Type
  "a plugin that provides turn-key comment feeds for your Movable Type weblog"
• Sun App Server Rebounds Significantly
  "+6% in a year is none too shabby for an 'also ran' app server"
• T1000 and T2000s blow away IBM, Dell in SPECjbb
  "most efficient platform ever for delivery of web and application services"

The first version of the Atom protocol isn't going to include administrative features, such as user management and weblog creation, but Jeff Blattman needed those features now. Jeff needed to be able to create users, weblogs and manage group blog membership remotely, so he created Roller Admin Protocol (or AAPP). He patterned his work after the Atom protocol so that, perhaps someday, it can be useful to the IETF Atom Working Group.

Jeff's did great work on the AAPP. He proposed it to the Roller dev list, designed the whole thing on the wiki and contributed all of the code and tests to the Roller project. He also wrote a complete set of AAPP docs for the Roller Developer Guide, created an AAPP SDK and finished some of my work for me by adding docs for Roller's Atom protocol implementation. Thanks Jeff!

Like Roller's Atom protocol implementation, AAPP is still considered an experimental feature  and is not for production use, likely to change, etc., but it is included the upcoming Roller 2.3 release (both Atom and AAPP will be disabled by default). Take a look and let us know how well Roller Admin Protocol will meet your blog server admin needs (direct feedback to the roller-dev mailing list).

Thanks to Simon Phipps and Tim Bray for putting together another blogger gathering at JavaOne. It's Thirsty Bear time again. Here are the details straight from Simon's blog:

Event:

JavaOne Bloggers' Social

Venue

Thirsty Bear Brewing Company

Date & Time

5:30pm-7pm, Tuesday, May 16 2006

Registration:

Please register!

I'm looking forward to meeting-up with some of the folks from the Roller-dev mailing list; sounds like Anil, Raible, Sean Gilligan and a couple others will be there. Roller bloggers, users and developers come on down (but don't forget to register).

Anne Thomas Manes has written an insightful overview of the REST vs. WS-* debate. She starts out by explaining that she doesn't just drink the WS-* Kool-Aid.

I'm one of the folks responsible for mixing the Kool-Aid. I presented at the W3C Workshop on Web Services (representing Sun). I participated in numerous standardization efforts at W3C, OASIS, WS-I, uddi.org, and JCP. I have a vested interest in making sure that WS-* succeeds.

She covers some of the same points I'm covering in my Atom talk (JavaOne 2006 TS-1756), which includes a brief overview of the debate, but she goes a bit further than  my slides dare to go with this:

But I can't ignore the debate between REST and WS-*. I'm a huge proponent of the KISS principle. So I don't recommend using WS-* for all service interactions. If an application doesn't require enterprisey infrastructure semantics, then it's much more appropriate to use a simpler middleware system, such as "plain old XML" (POX) over HTTP. In fact, for applications that require Internet scalability (e.g., mass consumer-oriented services), POX is a much better solution than WS-*.

Sounds like she mixes the Kool-Aid, but she stopped drinking it some time ago.

Anne was responding to a blog post by Mike Herrick, which compared big-and-bloated WS-* to big-and-bloated J2EE. I wonder, now that J2EE has been reborn as JEE5 and EJB as we once knew it is dead, does WS-* need the same treatment? Anne's post seems to imply that WS-* just needs better tools. And, does simply making it possible to build and consume RESTful web services with a WS-* stack, as JEE5 does, go far enough? Maybe I'll get some answers in Mark Hadley's talk on building RESTful Web Services with JAX-WS (JavaOne 2006 TS-1222).

Tags: REST, Atom, Java, JavaOne2006

With the book out of my hands again, I'm trying to return to normal life by doing things like leaving the house, watching movies and playing games with the kids. Tonight's fun was The Awful Green Things from Outer Space, an easy-to-play and very cute Steve Jackson mini-game (a board game) with an interesting set of rules. It's perfect for Alex, who at 9 is absolutely fascinated by game systems and rules.

TAGTFOS is about a space ship that is boarded by green monsters; monsters who grow each turn from eggs to babies and from babies to egg-laying adults. To fight off the monsters, the ship's crew members grab whatever weapons they can find, things like bottles of acid, cans of rocket fuel and blow torches. Problem is, the crew has no idea what the weapons do until they use them (and they change every time you play). So, if you fire a stun pistol into a room full of monsters you might blow 'em to bits, which sounds good until you realize that all of those bits will grow into babies next turn.

I got my copy of TAGTFOS from The Dragon magazine back in '79. The game is out of print, but you can find copies at used game stores online (around $15 a copy).

We're testing release candidates of Roller 2.3, so expect a release in next week or two. This release is a big one because we skipped 2.2, so we've got twice the normal monthly dose of features. Plus, we're shooting for full Apache license compliance. Read on for more [Read More]

Simon Phipps: We've gathered bloggers for free beer each year at JavaOne for the last few years - what say we do it again? I propose 5:30pm-7:30pm on Tuesday 16th May. Anyone interested? If so, say in the comments I'll go try to book the usual space! I can promise at least two well-known Sun bloggers will show up apart from me...

This will be the 3rd annual meetup. Great things happen at these blog meetups, BTW. I wouldn't miss it for the world.

Andy Oliver: Spent much of the day on something stupid that I did wrong. Thank god for the NetBeans's WORKING debugger (hear that Eclipsers...their debugger works AND they don't eat more CPU and memory than every Microsoft app I can get running on a Mac including Virtual PC combined -- simultaneously!!) -- or I might have never found it.

RTP blogger and JBoss hacker extraordinare is diggin' Netbeans. Not sure how long his Netbeans honeymoon will last; Andy's a tough customer.