A Brief History of Harmony

[This post represents my own memories of my own experiences. All stories have multiple sides, multiple perspectives, multiple angles, so if you’re interested in the story you should talk to other people involved in this first year of Harmony.]

I first heard about Project Harmony around the end of May last year when Amanda Brock sent a message to the FLOSS Foundations mailing list, inviting people to join some face-to-face and phone meetings around FOSS contributor agreements. Not a whole lot of detail, but the general gist was “Wouldn’t it be nice if we could do something a little more like Creative Commons, so projects don’t have to spend so much time creating contributor agreements, developers don’t have to spend so much time to understand them, and lawyers don’t have to spend so much time figuring out if they can approve employee’s requests to get them signed so they can contribute?” This got my attention. I’ve been a long-standing advocate of easy-to-understand legal language, a position I’ve held firmly through the drafting of the Artistic 2.0 license and the Perl contributor agreement, and through my participation on one of the GPLv3 review committees.

At the beginning of Harmony, I wasn’t involved with Canonical, though I’ve been an Ubuntu user and packager almost from the beginning of the Ubuntu project, and a number of my friends work at Canonical. I didn’t know Amanda, didn’t even remember hearing her name before. But, Harmony looked like it had interesting potential, and that was enough for me.

The first few meetings struck me as a little odd. They felt like political grandstanding, a lot of flash with little substance. It reminded me of the GPLv3 review committee phone calls. I kicked in ideas–I can debate like a lawyer even though I’m not one–but I didn’t have the sense that it was going anywhere. Looking back, that was an important first stage. It was how we all got to know each other, where we all stood, what was important to each of us, what we were there to support and to protect. It was a process of embracing chaos.

The group decided to adopt Chatham House Rule for our discussions. I’d heard of it before, but never actually used it. At first glance it seems quite sensible: encourage open participation by being careful about what you share publicly. But, after almost a year of working under it, I have to say I’m not a big fan. It’s really quite awkward sometimes figuring out what you can and can’t say publicly. I’m trying to follow it in this post, but I’ve probably missed in spots. The simple rule is tricky to apply.

At the second face-to-face meeting in early July, the massive block of email CCs that we’d been using to communicate was replaced by a mailing list. Someone brought a discussion piece to the meeting (no name, because of Chatham, but it’s already public knowledge that they were affiliated with SFLC, to my knowledge participating in Harmony as an interested volunteer). It was explicitly not intended as a first draft, but it did read very much like a contributor agreement, though rather too dense. The idea of drafting in options started to take hold. The sense was that we, the Harmony community, were not any kind of authority to say what was “right” or “wrong” in FOSS legal strategies. We were, instead, a rich collection of experiences in FOSS law and could help by explaining what choices FOSS projects could make, and the reasons they might pick one alternative or another. The agreements would then reflect the best practices around the various alternatives, basically “if you make this choice, here’s what experience has shown is the best way to do it”.

Through the rest of July, I was completely absorbed by OSCON planning. This blackout-month before the conference has been pretty common for me every year since I took on OSCON, and talking to other conference organizers I hear it’s common for all of us. In August I took a new job at Canonical and dove straight into “drinking from the firehose” to increase breadth and depth on my knowledge of the distro architecture. I also moved from the UK back to the US, suddenly and unexpectedly, as a visa requirement for taking the job. All that is to say, I was quite busy for a few months, and as a side-effect I didn’t really track what was happening in Harmony. I made it to a phone call or two, a face-to-face meeting in Boston, and half-way followed the mailing list, but I was mostly out of it. And, if I’m being completely honest with myself, I have to say I pretty much lost interest in Harmony during those months. I haven’t made any secret of the fact that I had a bad experience in the GPLv3 process. As a committee, we worked hard thinking through the issues and making suggestions, and as far as I can tell none of those suggestions for clarification, simplification, corrections, or perspectives on the needs of the free software community made it into the final draft. The final text released as the GPLv3 is, to me anyway, entirely disappointing. I saw Harmony going the same way. Those first drafts, intended as a seed for discussion, became the real drafts. They were bloated and dense legalese, and even though we had healthy discussions in meetings or on the mailing lists, somehow they didn’t seem to have any visible impact on the documents.  In my mind, and it was probably entirely unfair, I figured that since I had a bad experience with the GPLv3 process, and was having a similar bad experience in Harmony, that the similarity probably wasn’t a complete coincidence, and might just be rooted in the drafter. I can’t really say for sure, but I do know that my suspicion killed my motivation to participate. I’ll call this the start of the “Dark Ages”.

At Linux Plumbers Conference in November, Michael Meeks gave an excellent keynote about LibreOffice, that I attended with interest because we’d already started talking about shipping LibreOffice instead of OpenOffice.org in the 11.04 release of Ubuntu. In the talk, he mentioned Project Harmony in an unfavorable light. I don’t remember the details, but it was something about companies pushing an agenda of copyright assignment. The comments really baffled me, they didn’t seem to fit at all with my experience of the process or the goals of any of the participants. Even after chatting with him over a group dinner (a fun evening, I found out his wife and I used to work for the same non-software non-profit organization), I still couldn’t reconcile the gap between what I saw and what he saw in Harmony. I was going to set it aside as “somebody else’s problem” when a friend who I deeply respect and admire took me aside at Plumbers, and expressed concern about Harmony. I promised that I’d look into it.

So, I started asking tough questions, and what I found was both better and worse than I expected. I found that no one at Canonical had a bizarre agenda to force copyright assignment on the world. I also found that Canonical had an interest in replacing their current contributor agreement with a Harmony one, and that “success” for them was measured in community-driven, community-approved, and community-adopted agreements. All good. I also found that Harmony was pretty much stalled, all meetings on hold, waiting on a draft with some changes requested by the Harmony group (substantial changes, but shouldn’t have taken terribly long). Not good.

In a message to the Harmony list just before the mid-December face-to-face meeting, someone (again, no name for Chatham) said that SFLC would no longer be drafting the Harmony agreements. The general gist was that someone (again no name) had a bigger idea for how to improve the current system of FOSS contributions and copyrights, and that Harmony wasn’t enough of a solution. No animosity, wishing us the best of luck, with some possibility that Harmony versions 2.0 or 3.0 might play a part in a longer-term vision. I genuinely believe there was no ill-will in their parting as drafters. I still don’t know the reason for the drafting delay. Maybe they just got swamped with other projects, they do offer pro-bono legal services to the free software community at large, and tend to be quite busy. It is what it is.

I have the feeling that Harmony could have fallen apart at that point. Stalled for something like two months, the drafters bailed, rumbles of public criticism, and nothing to show for 6 months of work. You might think the tone of that face-to-face meeting would be a little dark. Instead, the universal reaction of the group was more like a head-scratching “Huh. That was odd. What next?” We had the seeds of something really good in Harmony, I had a deep sense this was true, and saw it reflected in the faces around me. We decided to reboot, set a schedule for release, go public, and pick a new drafter. A few participants said they’d find out if their companies could help fund the drafting work so we could actually hit the planned April 7th Alpha release date. I’ll call this the “Renaissance”.

In January, we launched a public website, the first incarnation of harmonyagreements.org. It was a single page containing the text from the current “What Is?” page, ugly as mud, and hosted on my own web servers.

Also in January, we started up weekly drafting calls. We took the earlier drafts as an inspiration, but started with a rewrite that I handed our drafter in January with the note that it might not be entirely legally accurate, but it was at least comprehensible to ordinary humans. And then we proceeded to pick that rewrite apart, line by line, word by word. The tone of these meetings was entirely different than the previous year. Partly, we all knew each other pretty well by then, knew each other’s values and priorities and could just settle down to work. Partly, the driving beat of the upcoming Alpha release helped us lock our attention on what to ship “now”. I worked hard to attend every meeting, even from exotic time-zones, on weirdly laggy voice connections. (I’m amazed at the patience other participants showed for including me in a conversation hampered by 2 second delays.) The discussions ranged over whether the drafts accurately reflected our intentions, had the intended legal effect, were understandable without a legal education, and would fit into the customs and cultures of existing FOSS projects. Every week our drafter brought a new version of the document (we worked on it as one big file with variations embedded), with substantial changes based on the group discussion. And every week we combed through those changes to make sure they were what we, as a group, wanted. People ask who drafted the agreements, and the answer is that we all did, together.

As we approached the April release date, we had a professional designer, Ben Pollock, prepare a proper design for the website (I highly recommend him, and hope he gets lots of business out of it, because he charged us beans). Oregon State University’s Open Source Lab agreed to host the site and our mailing list (they charged us nothing, please donate). I had a frantic last few days getting everything set up for the launch. To give geek credit where it’s due, the site is static generated HTML (anything else seemed like overkill for the handful of pages, we’ll expand later) with templates processed using Template::Toolkit (the libtemplate-perl package in Ubuntu), and both templates and generated pages stored in Git, so my “site launch” is a simple ‘git pull’. The comment processor for the mailing list that generates the comment review page is a Python script, basically a small finite-state machine using line-by-line pattern matching. (I’ll have to see if I can use Ruby or PHP for the agreement picker forms on our 1.0 website, for even greater compulinguistic diversity.)

On launch day, I woke up early to set the new website live just before Amanda’s panel at the European Legal Network Conference in Amsterdam. I don’t see the names of everyone who spoke at that panel posted publicly, so I won’t list them here. But, as far as I know the ELN conference is completely open to the public, so someone will eventually post it. After resolving a last-minute technical crisis with the review mailing list, and answering some press questions on the launch, I headed over to the Linux Collaboration Summit for the day. Bradley Kuhn slipped me in as a surprise special guest at the end of his Legal track, where I walked through some very last-minute slides, and a round of questions. The questions were good and thoughtful, captured well by Jon Corbet in his LWN.net post on the session. My immediate action items from the feedback in that session are around transparency: to replace the old “temporary” mailing list for drafting work with a second publicly archived list hosted by OSU OSL, and to put up a Sponsors page on the Harmony website as soon as we get approval from the various donors (names again withheld for Chatham). To anyone who is concerned about unfair influence by our drafter (even after I described how our drafting process actually works), let me say that sponsorship for the drafting work was only promised to the Alpha release, and is unlikely to be continued, since the workload of integrating revisions from the public review process is light enough to be handled by the volunteer participants.

I’m proud of the community group we’ve built around Harmony. I don’t think there has ever been such large-scale harmonious collaboration around FOSS legal strategies before. I take it as a positive sign for the future, that the FOSS legal community can embrace diversity, work together, and change the world for the better. I’m pleased with the template documents we’ve produced. They aren’t perfect, certainly, but they are a good reflection of current best practices, and I hope that through the public review process they’ll be even better by the time we make the 1.0 release.