Probably very few people know this, but when I first heard of the idea of doing what has become known as Callisto, I hated the idea. In fact, I attempted an ultimately unsuccessful argument against it. There were various reasons for this, but chief amongst them was that I didn’t think it would succeed.
I was wrong.
Callisto has already been a major success for the Eclipse community. Shipping 10 projects on the same day is no small feat. Everyone who worked on making Callisto a reality deserves a huge round of applause for a job well done. And I need to make this really clear: the idea did not come from the Foundation, and the credit for its success is clearly due to the projects and committers that came together to make it happen. In particular, I would like to single out David Williams from the Web Tools Platform PMC who put a lot of time and effort into the peer leadership required to make Callisto a success.
That said, I think that the real impact of Callisto is yet to come. That’s because cultural change always occurs slowly, and it is the cultural impact of doing annual release trains which I think will be the most impressive over time.
The first major cultural impact will be on the committer community and the project development process within Eclipse. Don’t forget that the release train idea was a bottoms-up phenomenon and that participation in each train is voluntary. Each project decides to join on their own. But the interesting side-effect is that more projects want to join the release train. It’s like a rite of passage: being capable of participating in the release train will come to mean that a project has grown up sufficiently to become a fully mature pr oject within Eclipse. This year, the Callisto projects had a 9-item list of requirements to participate (see the Callisto page under “Requirements”). I’m predicting that next year will have a longer list. But working with a common set of requirements and expectations will have a profound effect on all of the projects on Eclipse as they grow and mature. Most tangibly, this is a process by which the original Eclipse project can instill in the newer Eclipse projects some elements of the Eclipse Way. Not by coercion or by preaching, but by leadership within a group whose membership is purely voluntary. I predict that this will have a profoundly positive impact on Eclipse by creating a more cohesive committer community over time.
The second major cultural impact will be on the technology consumers who use Eclipse. For years the software industry has been conditioned to accept and to even expect failure. Microsoft is a particularly hapless victim of this, to the degree that it has even spawned a cottage industry around explaining the Broken Windows Theory. Contrast that with the fact that for now three years in a row, the Eclipse project has shipped on June 28 [1]. And now with the release train, 9 additional Eclipse projects are shipping on the same day. I believe that this degree of predictability will have a profoundly positive impact on the Eclipse ecosystem, as consumers (a) build their plans with a high degree of confidence that their underlying technology will be available on schedule and (b) aspire to build development plans which have the same degree of predictability.
I really don’t think that any of this happened by accident. Accordingly, I would like to recognize the efforts of John Wiegand, Kevin Haaland and John Kellerman who persevered in overcoming my original objections to doing Callisto. Eclipse is a better place because of their foresight.
I am sure that others can think of additional cultural implications that annual release trains may have on the Eclipse community. Comment away!
[1] This year, the projects are going to be done Callisto on June 28, but the Foundation needs 48 hours to ensure that the mirrors have been populated before turning on the download taps. You can start downloading Eclipse 3.2 and the rest of Callisto on June 30.
