JBoss AS 5.0 - Changes To Come
November 30th, 2008 Radim Marek
Together with other bloggers I’ve asked earlier this year the question about the release plan for the long awaited JBoss AS 5.0. Five months later, I’m happy to say, everything suggests the big moment is approaching fast.
Let’s review what we’ve got so far. Available for download is the second candidate release. Next in line, and probably most important is the information coming from the Application Server project lead, Dimitris Andreadis, who confirmed the release date to be December 2008. Another piece of the puzzle is the fact this week the JBoss Microcontainer, core component backing new version’s architecture, reached general availability release 2.0.0. With the rest of the projects being now quite stable, it’s easy to see ‘it’s happening’. Hurray! Or not? Let’s take another look.
I guess the impression from all the projects I’m using is that release candidate is almost final product. Unfortunately during testing of both first and second release candidate of JBoss AS, and now working with regularly updated trunk, the impression is quite different. Broken deployments with even quite regular WAR files, introduction of minor but important changes, and general integration problems don’t give me much of assurance. Having closely watched the process behind introduction of Glassfish v3 Prelude, I have to say I won’t be so wrong to say these two processes were significantly different. Purely subjective test, but one that will form first impression of many users, is to migrate the existing projects. Whereas even the migration to Glassfish (v2 and v3) was relatively straightforward, and I was able to use v3 Prelude without almost any changes long before it reached GA, the same task for JBoss AS 5.0 is pretty tedious. Not saying impossible, but even as a experienced JBoss user, I can’t help myself from feeling a bit frustrated.
Before making any judgement, there’s an important twist. Business - the always first and foremost root of all twists. The release of JBoss AS 5.0 is a matter of the community product only. Hand in hand with restructured and updated jboss.org, it’s important to notice it won’t be supported. Therefore it will take another 6 to 8 months (my rough estimate) to release EAP 5. In that time we might even see JBoss AS 5.1. With this in mind I’m not sure how significant is the December release for all of us - the developers making their living out of it. Do you really want to stay with the bleeding edge technology and experience all teething problems? You have to answer this.
Does this situation make me feel sad? Yes, it does. As I said before, this version is expected for such a long time, the new release will be significant and change the way how new applications are developed. But it has already left a slight bitter taste in my mouth. This year I’ve heard bits and pieces from different sources on the theory how such a open source project should be made commercially more viable, especially by loosening control over the community side of things. So I guess that is it. Six month ago I was asking question why should I use EAP if I ever start-up my own company, and today I’ve got answer.
Ignoring the the ground-breaking technological changes in version 5.0, I’m pretty confident this release will bring change in the way how open source projects are commercialized. RedHat needs to make money, no question about that. I just wonder what kind of impact this transition will bring to the users who made the original user base. In mean time, I’m going to enjoying standalone technologies, which make the JBoss portfolio so tempting - Seam, Messaging, Remoting, Cache, AOP and newly Microcontainer.
Posted in jboss | No Comments »