Chairing the EclipseCon Program Committee

First, I have to warn you that if Bjorn Freeman-Benson ever approaches you and says, "Hey, wanna chair EclipseCon? It really won't be much work..." you should turn and run. Fast.

Seriously, it has been a lot of work, but also very rewarding. Bringing a program together is challenging in many ways - at its heart, it's a fundamentally large bit of secretarial work, because there are many submissions, different categories, possible tradeoffs (more short talks and fewer long talks?), and a myriad of small issues (such as not putting two talks with the same likely demographic at the same time).

The program committee also has to look at the balance of the program, which is challenging in other ways - you don't want all the speakers to come from one company, or just from a few projects, or to just represent developers. We looked at different constituencies (committers, users, member companies, IT decision makers, etc.) and tried to ensure that some portion of the conference program would appeal to each of them. We also looked at a balance of old, current, and new technologies, because not everyone has been to conference before or is new this year. Finally, we tried to ensure that a few "virtual tracks", this year focused on RCP and embedded tooling, provided a credible draw for conference goers interested in those topics.

Apart from the sheer difficulty of reading and discussing so many submissions (and it's still going on - we're currently working on the short talks), one aspect of the program committee that proved challenging was trying to increase community involvement. This year we experimented with using a Bugzilla-like approach to submissions, which included allowing anyone to comment and even to vote on their favorite talks, including asking questions of the submitters. Ideally we would like to continue empowering the community, including the ability to rate comments by others and search more easily by adding additional submitter (and maybe community-provided) metadata in the submission form.

I'm looking forward to EclipseCon (ok, in part because my job as program chair will be over ;-) and hope to see you there - I'll be on an open source contributions panel discussion, the final plenary session, doing opening remarks one morning, and participating in the WTP events. If the abstracts are any indication, there will be plenty to learn about!

Onward and upward - M4 released!

Three interesting bits of news to report.

First, WTP's first step toward the 1.5 release and the Callisto train has occured. Go to the M4 down page for a version of WTP similar to 1.0 (with a few extra bug fixes thrown in; technically this is also a 1.0.1 I-build) but that works with M4 of Eclipse 3.2 and the other WTP prereqs.

Second, we've fixed some perf issues reported against 1.0, some of which represented regressions late in the release cycle. If you've experienced these and are feeling the pain, also check out the link above for fixes. (And if you still have perf issues, please report them - performance is a key deliverable for the 1.0.1 service pack.)

Finally, it's official: the Dali project, a technology project representing the merger of the former JSR 220 and EJB 3.0 technology projets, is here. This is great news for the Eclipse community and WTP in particular, because it means some of the strongest supporters of these technologies (Oracle, Versant, Sybase, Solarmetric, JBoss, and others) are all working together in a coordinated fashion on a single technology project. When these technologies mature, they will graduate into WTP and DTP as appropriate. See the Dali site for more details.

All aboard the Callisto release train

I've been doing a lot of press interviews this week, and everyone wants to know about Callisto - the code name for the joint release train containing Eclipse platform 3.2, WTP, TPTP, BIRT, and other projects slated to come out in late June 2006.

In some ways, there's nothing new here, except that a few dates moved and we agreed to a (modestly) less confusing nomenclature on milestone numbers, release candidates, etc. And put in place some processes to help projects get bug fixes from one another in a timely manner. So, no big deal.

In other ways, though, it's a huge deal: it helps users get a richer set of product features in a straightforward fashion, without having to worry about the "recipe" - download version X of Y, version W of Z, etc. and hope that when it's all over they have a working Eclipse. And by agreeing on milestone dates, we make it possible for adopters to more easily test their own products.

The most important part of all, in my mind, is the testing and exposure that will come from people exercising *more than one* project at a time. I think this has huge potential to improve Eclipse's UI consistency and ease of use over the long haul. (In the short term, think Bugzilla.) I'm also excited because some of the projects have - and apologies for the marketese term - cool synergy. For instance, using TPTP to monitor the performance of J2EE applications developed with WTP, then using BIRT to produce readable test reports. That's possible today, but by making it easier to get all those projects together, I think we'll start to see more people engaging in cross project actitivies and therefore driving the projects to work better together.

WTP's first steps in boarding the train will happen in mid January, when we're planning an initial milestone whose sole goal is to get our 1.0 codeline building and running successfully on the latest milestone of the platform. At this stage, we won't be adopting new platform APIs, just making minimal changes. The next step is to adapt to those APIs as part of M5. The M5 Callisto milestone covers a two week period - the platform goes first, on 2/17, with downstream projects following suite and completing the milestone release on 3/3. All the Callisto projects should be aboard the train in this milestone, so users and commerical adopters should be able to start looking at the result. We'll be using an update site (details still TBD) to let people sync to the non-platform pieces of the Callisto release easily.

Lots of challenges remain (and I wasn't exactly needing another weekly meeting on my calendar), but I'm very excited by what we're trying to produce.

WTP 1.0...the last mile

It's been a long haul, but we're finally about to release the "1.0" version of the Eclipse Web Tools Platform project. Technically we're still in cool down - the final window for adopters to raise issues - but the hope is that we can declare RC5 the final set of bits.

For those of you who are new to the project, Web Tools (or "WTP" as we invariably shorthand it) contains editors, wizards, tools, and APIs to help Eclipse users develop Web and J2EE applications - from HTML editing to JSP debugging to Web Service exploration and a lot of points in between. It's useful right out of the box, but also designed for companies to incorporate as part of their extended commercial offerings (which is how I manage to keep food on the table ;-).

WTP is a great project, but it's also a *huge* project - lots of code, lots of features, lots of tools - and managing all those moving pieces to a single release date is a major challenge. We know this release will have all the classic "1.0" ills, but it's still a tremendous accomplishment, and one that we're happy to get out the door.

Now that we're done with the code itself, we're working hard to get the message out, in both formal (read: press release) ways and informally via blogs (shown here), postings, etc.

One of the common questions that comes up is "what's next?". In the short term, the answer is "1.0.1" - we're scheduling a quick turnaround service pack for February 16th, built on top of the Eclipse platform's 3.1.2 SP.

After that, the next release is WTP 1.5, which is part of this cool effort called "Callisto": A coordinated release of major Eclipse projects (platform, DTP, WTP, TPTP, BIRT, selected tools and GMF) all on the same day in late June of 2006. This will be a first for us as project leads, so wish us luck with the coordination effort it's going to require!

In Callisto/WTP 1.5 part of our focus will be on adopting new architecture and API changes in version 3.2 of the platform, which forms the "engine" for the Callisto release train. After that, we'll be looking to do more aggressive feature releases in a subsequent WTP release...perhaps numbered 2.0?

Another frequent question is: "What's changed since 0.7?" The biggest conceptual change is the shift from provisional to real APIs. We've also spent a lot of time on stability and bug fixes, introduced a new runtime functionality modeling technique called "facets", beefed up the docs, and aligned our project model more closely with the Eclipse platform.

Do us a favor and give RC5 (soon to be1.0 GA) a whirl, and report any issues you find using Bugzilla. You can also post questions or engage in discussions with the WTP community using the newsgroup.

I'll be blogging on WTP, BEA's Eclipse-related activities, and other aspects of my professional life. Stay tuned for more...