Skip to main content

Exploring the Java Research License

June 24, 2003




One of the java.net-related announcements at this year's JavaOne conference was the Java Research License. The license is intended for those wanting to use software licensed under SCSL (the Sun Community Software License) for research purposes. We invited several java.net bloggers to discuss the license in an email exchange. Here are the highlights of their discussion.

The participants were Danese Cooper, John Mitchell, James Duncan Davidson, Jonathan Simon, and Jason Hunter.
For those not used to some of the popular acronyms used in such discussions, IMHO is short for "in my humble opinion," and IANAL should be read "I am not a lawyer."

This article is intended as a conversation starter and not as a final word. Please feel free to add your comments to this article. Of course, you'll need to register for java.net first :-)

Lawyers and Click-Through

James Duncan Davidson kicked off one thread in the discussion focused on section III of the JRL and the requirement of a click-through license for works released under the terms of this license.

James: I was hopeful when I first read the FAQ for the license. It seemed like decent non-commercial license. However, when I actually read the license itself, several problems emerged. And when you read the FAQ the right way, the problems are in there as well. For example, as you'd expect for a research license, the FAQ states:

"You may publish your work in the usually accepted academic manner as long as you reference the Java Research License and include the correct copyright information."

Seems nice enough. Researchers can share freely their ideas. Except for the fact that it's a click-through license, which means that everyone who wants to share has to go to Sun's site. The FAQ says:

"For enforceability, Sun requires a click-through license."

So first Sun says that you can share. But then you can only share if you go through them. If I do some research and then go to a conference of my peers to present the research, according to this, not only do they have to agree to the terms of the license (probably not a big deal), but they have to go to Sun to become a licensee by clicking through (a bigger deal).

Annoying. A research license should be a non-commercial use license that is required to be redistributed with any modifications. End of story. Yes, I know that the legal enforceability of the license is strengthened by having the centralized authority. But the current rules are a great impediment to its use, and only increase the perception that Sun wants to maintain a stranglehold.

John: I don't read JRL Section III.A quite that way.

I do strongly concur that the click-through requirement for sharing the
source code (III.A.2) is onerous and annoying.

III.A.3 covers executable code that only requires the use of a compatible
license. That licensing compatibility does not require either a
click-through or that the receiver be a licensee.

III.A.4 covers talks, papers, books, etc. and does not require the
receivers to be licensees, but does strongly limit the amount of direct
exposure of the technology to relevant excerpts.

As always, I'm not a lawyer. Always seek competent professional advice.

James: It's confusing. And that's its downfall. It's more confusing than many other licenses. I really wish that the courts would enforce readability of licenses by laypeople.

John: To be clear and fair to Sun's lawyers, at this point in time and given the existing legal precedence, that sort of click-through mandate is the only reasonable way to establish a sufficient basis of diligence by Sun in
protecting Sun's intellectual property.

James: Maybe. I'm not so giving to Sun's lawyers. Has click-through of a license been tested in court? Most software license strategies have not, to the best of my knowledge. The Creative Commons license set was authored by some of the very best legal license minds for use by anybody with deep or shallow pockets. They don't require click-through. I'm not convinced. And even if it is thought to improve the chances of enforceability, it's like the argument that you improve your life expectancy by never traveling. It's an impediment to use, and every impediment means that less people are likely to use it, which limits the benefit gained by pursuing the effort on Sun's part in the first place.

John: If nothing else, the click-through shows some level of intent. Showing intent makes a big difference when you're trying to recover damages.

Published Papers

Jonathan Simon struck up a side conversation on the restrictions of the JRL that might apply to published papers that include materials covered by the JRL.

Jonathan: If I publish a paper on some J2SE research,
will all of my readers have to do the click-through license? I don't like the idea
of having to agree to a license to read an article.

John: Again, I don't read the license that way.

They only have to go through the click-through if they want to get, for instance,
your changes to the Materials. In this case, the word "Materials" has the specific meaning as given in the license.

For a paper, book, talk, etc., you are free to judiciously excerpt portions
of the Materials as long as the aggregate doesn't constitute "too much."

Danese: FYI, Sun has long allowed code samples to be published in qualified research journals with prior written permission by Sun. Not listed in the terms of this license, but FYI can usually be accommodated.

Production Usage

The final discussion was sparked by Jason Hunter's thoughts on the following excerpt from the JRL FAQ (frequently asked questions).

5. When do I need to get a commercial license?

This research license is only for initial research and development projects. If you decide to use your project internally for a productive use, and/or distribute your product to others, you must sign a commercial agreement and meet the Java compatibility requirements.

9. Am I required to keep my research "Java compatible?"

No. The license encourages you to innovate and experiment using the Java technology core. It is expected that research implementations will not be compatible with the Reference Implementation from Sun. This is OK.

Jason: It seems ludicrous to me that if you use the code internally for production use you must meet any compatibility requirements. If it's fully internal, compatibility with what's external should be wholly unimportant.

What good is a research license that doesn't let you actually use the results of your innovative experiment in any productive manner?

Jonathan: I can understand non-distribution to an extent. But what
if I build an in-house, custom XXX, I can't use it. That's just bizarre.

John: IMNSHO, that's a FAQ about cash -- i.e., Sun wants the cash and the ability to tout that they support "open" research. It's exactly the same as calling the JCP an "open" process -- a lie.

To be clear, it's Sun's IP, so they should be allowed to do what they want
with it, but ... I think it is, at best, misleading how they market this
stuff trying to leverage off of the backs of people who don't know any
better (caveat usor!) to people who don't know any better (caveat
emptor!).

Of course, there's also the question of enforceability. :-)

James: Yes, I've never begrudged the fact that Sun can do with their IP as they will. I hate that sometimes the truth gets bent in a desire to appear one way and not the other.

Danese: I think I've heard it explained thusly:

The problem is not isolated internal deployment, its productive deployment that leaks outside of what Sun would consider internal boundaries. For instance, is it okay to give to a partner working on the same project? How about a subsidiary? How about a sister university? IANAL, but I think Sun Legal probably can't see a bright line that bounds such deployments without endless corner cases.

James:
Right. But there's always some corner case. Sun's optimizing out the corner cases also optimizes out the area conducive to use.

Jason: Does Sun ever sell a "company-wide" software license? I'm sure it does. The same rules can apply. If they wanted to solve this problem, if they were motivated to solve this problem, it would be easy to solve.

Summary

Looking at the license as a whole, the group did not view the JRL as much of a change in policy.

Danese: IMHO (and IANAL) the JRL doesn't actually represent much of a change of terms from what the research and academic community could do under SCSL (there are some small changes around export), but it does clear away all the language in SCSL that is confusing, if you are only planning to engage in research. We got feedback that, with SCSL, it was too hard to parse what your obligations and rights were if you just wanted to do research. The JRL tells you that, and that was its primary purpose. If you want to make non-research use of your work however, you're back to SCSL terms.

John: I concur on all points. The JRL is a letdown.

FWIW, I personally think that we're all on the side that wants Java to be successful. That's part of what makes many of the things that Sun does to/with Java so frustrating to us.

Daniel H. Steinberg runs dimsumthinking.com and is former editor-in-chief of java.net.
Related Topics >> Education   |   Research   |