Why is the Jini Bottled Up?by David Sims
What's the biggest obstacle holding back Jini, Sun Microsystems' technology that aims to let devices connect smoothly to networks?
Is it the security problems that must be addressed before they can be fixed? Is it the lack of a highly visible killer app? Or is it that there's no grass roots momentum among young, independent hackers, the fact that interest is centered in big companies?
Third Jini Community Meeting|
• Jini developers met in Santa Clara on Tuesday, March 28, 2000, at the O'Reilly Conference on Java.
• Two previous meetings took place in Annapolis, MD., last October and in Aspen last May.
• An ongoing goal of Jini developers within and outside of Sun is to build a community organization strong enough to take over the stewardship of the technology from Sun.
None of these. The consensus at the Third Jini Community Meeting in Santa Clara on Wednesday, March 29, 2000, was that the biggest obstacle to Sun's Jini could be the terms of the Sun Community Source License (SCSL). While developers are keen on Jini's potential, they can't always sell their companies on the notion of contributing intellectual property and donating services and time to develop a technology that Sun may decide to reclaim.
"I personally think that Jini needs to belong to no one, or to an independent standards body, or you're just not going to get any traction," Tim O'Reilly told the developers. "You've got to get it out there as an open standard, and an open standard is not a Sun standard. Some of it is fear of Sun, but some of it is just friction, the hassle that you have to figure it out."
As an example, he asked them to think about how the Web would differ if Cisco owned TCP/IP and anyone who used the protocol had to pay a small licensing fee to Cisco. (That could well make Cisco the most valuable company in the world -- nevermind.)
But what is free? Dick Gabriel, the Sun Distinguished Engineer who has been among the strongest advocates for Sun giving Jini to the community, asked, "Is it free as in free beer, or free as in free speech?" In other words, is it the cost of licensing Jini, or the fact that it has to be licensed?
Definitely the latter. Developers agreed it was Sun's hooks in Jini and the hassles of the license, rather than any license expense.
One developer, from a large copier company, said his company wasn't into him developing code that would then be released to a community. "You just can't put free code out there, within our environment. " Another agreed that releasing IP was tough: "If you have a big patent treasure chest, as a company you're going to do well. If you give it away, you're not."
Still an overwelming consensus among the developers is that it's not yet time for Sun to release the source code to a Jini community -- since the community isn't ready for it. Sun's licensing representative, Danese Cooper, suggested that if the group of developers was more officially organized, with a charter, that it would be time to consider releasing the source.
Handing over the reins
The meeting was characterized by the frank discussion over Sun's ownership, and by the long, painful birth of a Technical Oversight Committee, which is to begin taking stewardship of the technology from Sun.
Dick Gabriel and fellow Sun engineer Ron Goldman, who have led the community building efforts, began the meeting by announcing they were stepping out of the role of guiding Jini to start a Laboratory for Business Engineering within Sun -- a place to learn, in Gabriel's words, "how to do development in open conditions." Jini is just one of Sun's efforts in this area; it's also opened Solaris.
"Our goal has been to create a collaborative community based on open source models," Gabriel said. "The most important shortcoming, is the failure to create a collaborative community."
To pick up that task, they handed the meeting over to Sun engineers Ken Arnold and Jim Hurley, who spent much of the rest of the day building consensus around the idea of electing a Technical Oversight Committee, a group that would act as stewards of the technology in line with Sun Microsystems' stated goal to let the community of developers guide the technology's development. Sun's Jini Architecture team has been the acting steward of the technology.
There was a lot of discussion about how to choose a Technical Oversight Committee, and what it should do. In the end, over the objections of a few vocal members, the group of 50 or so Jini developers chose a team of nine people as the new committee, with a mandate to figure out the scope of what they do, and how long they should do it.
Some developers wanted to give Sun a great deal of power, to let it show "some leadership." Others were skeptical of giving Sun too much control over a technology that Sun is actively marketing as following the development models of open source. Why should they spend their development time improving a technology that Sun may ultimately decide to reclaim?
This meeting was smaller than the other two -- why?
Outcome of meeting|
• Nine people were nominated to be on the Technical Oversight Committee.
• Ron Goldman and Dick Gabriel, who helped get the community organization rolling, are stepping aside to build a lab for open technology development within Sun. Ken Arnold and Jim Hurley will shepherd the community.
One developer, who identified his employer as a large, East Coast telco, thought it was for two reasons. First, he identified some malaise around Jini in the public. Why hasn't it taken off? Second, he noted that in a democracy, when no one is being oppressed, there's very little motivation to participate in government.
To which Ken Arnold quickly retorted, "So we should oppress some people?"
No grass roots, no killer app
Security remains a big problem. Some developers said they show Jini to clients, internal or external, only to hear, "Great technology ... too bad I can't use it."
Another obstacle may be that, as one developer noted, there's no grass roots interest in Jini outside of corporations. "Until we get 19 year olds interested in it, it's not going to take off," he said. There were several comparisons to Linux, which had the advantage of playing against the foil of Microsoft's operating systems. Jini also has a competitor from Redmond, Microsoft's Universal Plug and Play -- but so far there's very little resentment over it.
Given that, what Jini may need is a killer app. O'Reilly tried to steer the group off its discussion on technical committees, suggesting it was spending effort in the wrong place (building a bureaucracy) when it should be building a killer app to grab the attention of business and media. The bureaucracy they build, he added, could wind up being a further barrier to entry of new coders to the Jini community.
Even so, the group must know that to wrest Jini from Sun, it will have to be fairly official and organized. Although O'Reilly may have disagreed with their methodology, he and the Jini developers seemed to have their eyes on the same prize: getting Jini out of the bottle.
Return to O'Reilly Network Hub.