New DivaBlog

Copyright Assignment

Just noticed David Berlind's blog on OpenOffice.org's copyright assignment policies. It gets a few things wrong...for instance the assertion that the SISSL grants copyright to Sun. In fact the thing that does that is the Joint Copyright Assignment, a separate document which must be signed and on file at Sun before a contribution can be accepted into the main OpenOffice.org source tree from outside of Sun.

The JCA is the "new, improved" agreement. OpenOffice.org actually started with the same copyright assignment form that the FSF uses, but since mostly OpenOffice.org engineering happens in Europe there was much concern in the community to deal with problems arising from the unilateral transfer of copyrights in the FSF's form. The JCA was drafted by our friend Cliff Allen, a member of the Sun Legal Team and I did the proof-reading myself. At the time it was drafted the OpenOffice.org community was very happy to have an agreement that didn't ask them to do things they couldn't legally do (such as transfer their "moral rights", a concept that isn't handled too well in the FSF form).

The article makes it sound as if Sun is doing something exotic with the JCA in OpenOffice.org, but in fact all the FSF projects, all Apache projects, and many other open source projects ask for similar rights. The Apache folks will point out that they only ask for a license to relicense, but its essentially for the same reasons. Why? Well, in the US in order to use copyright law to "defend" a codebase you have to be able to represent a majority (51%) of the copyright holders. Imagine trying to assemble 51% of the hundreds of people who have signed a JCA for OpenOffice.org in a courtroom! Also, having the JCA allows Sun to maintain the dual-licensing (so if you donate code back under SISSL, Sun can also publish it under LGPL).

Licensing stuff is complex and hard to explain in sound-byte format, unfortunately...so its not surprising that David didn't get it exactly right. But he does have a point about the ability to respond to things like relicensing needs. I'm sure my pal Larry Rosen will want to chime in here, because he disagrees with the whole copyright aggregation idea. But Mozilla.org doesn't aggregate copyrights, and it took them a long time to track down enough of the contributors to get them to agree to add the LGPL to their license portfolio. No big company is going to unilaterally relicense. They'd risk a fork if they did. But the ability to build consensus within a community of current contributors without having to track down every contributor ever (even the dead ones) is compelling.

My apologies to Mozilla...

Memo to self:..."Don't piss off your friends un-necessarily!"

A couple of week s ago OSI released a document about License Proliferation. The thrash on License-Discuss has ranged from support to confusion to fear and loathing. There has been speculation about what we meant by "assymetrical" in describing the Mozilla Public License model. There have been outraged calls for OSI Board to apologize to the Mozilla community for saying "Mozilla is a failed experiment"...and of course assertions that OSI has declared war on corporate licensing, or on Mozilla or on CDDL or on everyone!

Okay, so first of all, I'm a Pacificist. I've never declared war on anyone, nor can I imagine myself declaring war. Try to remember this is software, not ethnic cleansing....

Second, I am a dolt; for not realizing that the Mozilla community would be upset by the sentence that has been parsed out of the document as the unfortunate headline "Mozilla is a failed experiment"...yes, the OSI Board should have caught the fact that we didn't refer explicitly to the licensing model. It would have been easy to avoid bad feelings. Especially since so many of us are great fans of Mozilla.org. I love Firefox. Some of my best friends are Mozilla committers or advisory board members. To quote the immortal Homer Simpson...Doh!

In fact, I'm one of the people who has consistently said that Mozilla the project is not a failed experiment.  I remember having that debate back in 1999 when Jamie Zawinski left in a huff.  Many Sun execs concluded back then that Mozilla was "over" and but I maintained and espoused a belief that Mozilla might surprise everyone someday.

So why the picking on Mozilla? From my point of view, OSI Board was just trying to point out in the document we released that the actual Mozilla license (the MPL) represents an important historical tradeoff on the part of OSI between the desire to build a deep repository of reuseable code and the desire to welcome corporations into the open source movement. We made that tradeoff despite our first goal to help promote recombinable open source code repositories. Moreover we wanted to point out that the MPL has caused some problems in today's Open Source ecosystem. These problems are not manifested within the Mozilla project. That's because Mozilla has successfully built a large and useful code repository under its project banner which is recombinant with itself. The problem is not with Mozilla project nor with Mozilla community.

The problem lies in two aspects of the MPL model when it is used outside of the Mozilla community: First that the license was not drafted for reuse and so encourages license proliferation by essentially requiring it...Second that the license explicitly builds a fence around its covered code which many read as a prohibition of code recombination outside of a single project umbrella. The result of this second aspect is that no single re-user of the MPL model has achieved a deep and recombinant codebase, and the Mozilla community itself eventually added the LGPL to its licensing strategy so that their code could recombine to the GPL community on some basis.

As many have pointed out within this debate, the GPL isn't particularly recombinant except with itself either, although the GPL is so widely employed (and doesn't require re-drafting to employ) that in essence there is a deep olympic-sized pool (or perhaps a small sea by now) of code that can be freely recombined under GPL. OSI acknowledges this goodness. But many corporate interests prefer the Mozilla model, at least as they are establishing communities, because (to paraphrase my former Sun colleague, Claire Giordano) ..."It fit our needs best".

So, what to do? Some members of OSI would like to deprecate the MPL model. I personally would prefer to see it fixed for reuse (as I've said before). What I specifically don't want to see happen is for the Mozilla folks to take our analysis of the problem as an attack. It's a statement. We all need to think about the issue of recombination of code vs. diversity of choice in F/OSS licensing. I remain hopeful that the tradeoff can still result in positive outcomes.

Postfacto Recognition

So, I just learnt via my friend Will Snow that a project I worked on...blogs.sun.com...just won a Chairman's award at Sun!

The Sea Refuses No River

Listening to my former bigger boss while sitting with Simon Phipps at OSBC. For those of you who have never been in the position of supporting an executive in your company through a speech, its a lot of fun. For one thing you get to find out which of the memes you've been planting inside the company have "taken" (you get to hear your words played back to you through someone else's brain). Its an interesting feeling.

Executive presos are typically built by a team; each person only contributes one or two slides. I know it seems bizarre to think about executive presos as a piece of collaborative performance art, but that's what they in fact are my experience (and not just a Sun).

Today Simon had the pleasure of seeing one of his slides played by Jonathan (it was the one about how the Amazon river is really a whole bunch of rivers) for those of you who were there).

Come to think of it, that picture can also be used to explain the way change is influenced in a big company...If the ideas the company carries forward in its messages are a river, you'd better believe there are a hundred tributaries and feeder streams that represent people behind the scenes.

While we're at it discussing Jonathan...he's still having trouble melding his views on the ongoing value of securing patents with the F/OSS view that the software patent system needs immediate and serious reform if not abolition. I would love to see Sun and other open source companies coming out more strongly in support of software patent reform than Jonathan is currently willing to.

Simon wants me to say here that as he actually sees it there are really two issues relating to patents and open source. One is that the system needs reform and the other is that until it is reformed (and perhaps even after that) it is vital to the open source community that corporations be held to the expectation that F/OSS communites be indemnified against fallout from the way the system currently works.

My favorite quote from Jonathan (which is perhaps cogent to Simon's point above)...Welcome to the Age of Participation

Expanding the OSI Board

Today is the first day of the new OSI year! We are installing five new board members today, including four individuals who live and work outside of North America. The new board is:

Michael Tiemann (President, pro-tem)

Danese Cooper (Secretary & Tresurer)

Ken Coar

Russ Nelson

Chris DiBona

Bruno Souza

Rishab Aiyer Ghosh

Sanjiva Weerawarana

Joichi Ito

I'm really excited to be starting this new OSI year! I think its going to be a great year for working on the challenging issues facing the greater open source community today.