Jini was a Sun invention, and when the company released the first Jini implementation in 1999, it soon started to organize a developer community around the technology. Jini provided a proving ground for Sun's future open-source developer community efforts, including java.net.

In the past seven years, the Jini community evolved from early adopters, to enterprises wanting to use Jini for business gain and, more recently, to an open-source developer community. The last evolutionary step was facilitated by Sun donating the Jini codebase to the Apache software foundation. Apache has accepted the project, and the logistics of the move are currently being worked out. While half the proposed committers are Sun employees, non-Sun contributors are expected to provide an increasing percentage of new code, as well directions for Jini's future evolution.

Dan Creswell is a committer on the Apache Jini project—the Apache project will likely have a different name—and was recipient of the 2005 Jini Community Contributor Award. Creswell, and Sun's Jini Community Manager Jim Hurley, spoke to Artima about the community meeting and Jini's future.

Hurley told us that about 75 people attended the meeting from thirteen countries, and that the meeting reflected a step in the evolution of the Jini community:

These [community meetings] reflect the growth and evolution of the community itself. We started the meetings in 1999. At that time, we were trying to understand what we were doing ... and were coming together to form a bond and a culture [around Jini].

Later, we got to a point in the community where we needed to make decisions on ... standards, and how we wanted to govern and manage ourselves. In the next phase of these get-togethers, we were talking about governance.

Then we got past that, and moved into a phase where we were talking about the users of the technology and about the projects [that built up around Jini]. In the sixth, seventh, and eighth meetings, there were a lot of presentations and discussions about how the technology was used, and what kinds of advances were there. We had presentations about commercial deployments and products based on Jini.. as well as integration with other technologies.

The collaborative projects have always been an important part of the community, and these meetings have been a great opportunity for people working on projects to get together and talk about them.

The next phase in the community meetings had bit of a different feel to me, because we got into the issue of talking about open-source licensing. The maturation of the community means that the community wants to have a bigger advancement in the technology, and also broaden the use of the technology. Licensing was a key part of that... Once we decided what license to choose, [the topic was about] how we do things in the open-source world.

The last meeting was also about a number of big changes under way as well. One of those is that we proposed Jini to the Apache Software Foundation. We decommissioned our old Jini.org collaborative Web site. Many of the projects on that site migrated to java.net, and we started a new site, a wiki [Editor's note: http://www.jini.org], which is an informational site about the things happening around Jini. And at this meeting, a lot of the discussions ... were about the community gaining a much more visible and active role in the technology.

Dan Creswell noted that not only did the discussion topics of the Jini community evolve, but so did the community's view of Sun's role in Jini's evolution:

To me, the last few meetings had a certain amount of frustration, because as I sat there, I watched a lot of people behaving like this technology is Sun's, and that ... Sun [is] responsible for it in terms of popularizing it, writing the code, doing the testing, providing finance for the [community meetings], and so on. There is still a little bit of that around. But there are signs now, at last, that the community is taking ownership of this thing for itself, moving forward.

One example is jini.org itself. In the past, it would have been a Sun-owned site, and it would have been [expected] that Sun provide editing and resources for it. These days, we have a growing number of people... editing those wiki pages, and expressing opinions. And that's new.

Sun's vehicle for releasing Jini was the Jini Starter Kit, a bundle of source code, binaries, documentation, and some utilities. The Starter Kit code will be the main focus of the Apache project, but Hurley noted that Jini has grown far beyond that initial Sun-supplied codebase:

The community still needs a "commons:" a basic set of code and standards that everyone could build on. There has to be a certain common set of agreements that, in my mind, were the Jini standards all along, and that's what we're proposing to Apache.

But... more code is being developed outside the core, and outside of what we're proposing for Apache. Whether these are projects on java.net, or projects maintained in other ways ... [all of those projects are] in Jini space as well, and are developed by the community at large.

When the open-source Jini project starts its life under the Apache umbrella, Creswell and his open-source collaborators will be in charge of the code base that formed the initial Starter Kit. Creswell noted that the community is still debating just what future versions of the Starter Kit would include:

[In the future], there will likely be a core chunk of code that will probably not be called the Starter Kit anymore, and then there will be chunks of code outside that core. And one of the decisions to make is who is responsible for which bits. Some people still think that everything should be in that core chunk of code. My opinion is that that would be a bad way to go.

The move to Apache has opened up Jini for use in any open-source or commercial project. However, Creswell noted that developers still need to learn more about the philosophy behind Jini to make better use of it. For instance, developers often think in terms of an enterprise application running in just a single VM, whereas the reality is that most applications operate in some sort of a distributed environment, which is what Jini was designed for.

What were your experiences with Jini so far? And what do you think it would take for the Apache Jini project to cause Jini to gain traction inside other open-source or commercial projects?