Distributions

How to enforce Debian's code of conduct

By Jonathan Corbet
September 10, 2014

The Debian Project, like many organizations in the free-software community, has put a focused effort into improving the environment in its communications channels and at its conferences. Among other things, the project adopted a code of conduct in April; that code is expressed in simple terms of desired behavior; examples include "Be respectful" and "Assume good faith." But Debian's code, like any code, risks being ignored if there is not some sort of enforcement mechanism behind it. According to at least one project member, an incident at the recently concluded DebConf event in Portland shows that the necessary enforcement is not yet in place.

The participants in the public conversation have carefully danced around the details of the specific incident — even the name of the offending speaker only slipped out later on — but the overall shape of the complaint has become clear from details that leaked out in various messages. Portland is Linus Torvalds's home town, so it is perhaps not surprising that he was invited to do a question-and-answer session at the conference. Linus, never known for mincing his words, was in typical form for this session — though, in truth, he was more restrained than he can be. He did, however, have some strong words about the Free Software Foundation; he is, it seems, still somewhat unhappy about how the whole GPLv3 drafting process went, and he let the audience know just how unhappy he was. Those who are interested can watch the video of the talk and reach their own conclusions.

It appears that Ian Jackson was not happy with this session, seeing it as a violation of the code of conduct. He duly filed a complaint with the project's anti-harassment address, seemingly requesting a statement from the project disclaiming Linus's words and a promise to not invite Linus to future events. The person receiving the complaint forwarded it to the Debian project leader on the reasoning that the anti-harassment team had no authority to make blanket statements for the project as a whole. In the end, it appears that the team reviewed the incident and concluded that no code-of-conduct violation had taken place, so no further action was taken.

After the conference, Ian went public with a message requesting "more guidance" for the anti-harassment team; he carefully avoided discussion of what had upset him. He would like to see the creation of a document for code enforcers starting with a number of examples of various types of undesirable behavior and the appropriate response. It should, Ian said, say who is specifically responsible for dealing with incidents, especially when they happen at conferences. He asked for an explicit rule that the Debian project leader should not be involved in the handling of complaints. Finally, he wants some guidelines describing what can be done about violations from guest speakers who may have already left the event (and thus cannot be kicked out) and discussion of "the rules for data sharing with other events".

Needless to say, this missive was not well received by some members of the community; it was seen as a call for censorship by some. So some members responded in ways that, in their own right, arguably failed to meet the code of conduct. One participant was banned from the mailing lists for his role in the discussion.

Others, though, took the request seriously, even though not all of them felt that a violation had happened in this case. Manoj Srivastava suggested that some sort of enforcement was necessary if the project actually means what it says about the environment it wants. Otherwise, he said, "we should strip language out of the CoC about being respectful to people and making attendees feel welcome, to avoid giving a false impression that those things are actually important and shall be enforced."

Steve Langasek, instead, argued that the project already has the tools and procedures it needs in place for dealing with incidents at conferences. The means by which complaints should be made is spelled out and the people responsible for dealing with complaints have been identified. With regard to the case where a violator has already left the conference, Steve did not think there was much to be done:

The stated purpose of the CoC is to ensure that our conference is a safe space for all members of the Debian community. In what way would a change in approach to dealing with a violation after the fact, where the offender is no longer at the conference, further that goal?

Didier Raboud said, instead, that, outside of severe incidents, the code should not be treated as a law to be enforced. Instead, it should be a guideline that all project members take to heart and try to help each other live up to.

In the end, there may only be one significant change from this event: the anti-harassment team seems likely to make a more concerted effort to communicate the results of a complaint back to the complainant. It appears that Ian felt he did not get a definitive answer back from the anti-harassment team; that did not help his overall feeling about how the issue was resolved.

The real truth of the matter is that the project's new code of conduct has not yet been seriously tested; the incident at DebConf Portland looks like more of a false alarm. As was seen at LCA 2011 or PyCon 2013, an incident that forces a high-profile enforcement action tends to leave a fair amount of controversy and bad feelings in its wake. Human nature being what it is, chances are that Debian will have to deal with such an incident sooner or later. So a discussion now of how the project should respond makes sense; it may help to avoid a rather less pleasant discussion in a future year. Even if the discussion appears to have few useful outcomes now, it should have succeeded in getting project members to think about how they want their code enforced.

Comments (48 posted)

Brief items

Gentoo Linux 20140826 - Iron Penguin Edition

Gentoo Linux has announced the release of 20140826 - Iron Penguin Edition. If you've thought about trying Gentoo, the Live DVD is good starting point. "The LiveDVD is available in two flavors: a hybrid x86/x86_64 version, and an x86_64 multi lib version. The livedvd-x86-amd64-32ul-20140826 version will work on 32-bit x86 or 64-bit x86_64. If your CPU architecture is x86, then boot with the default gentoo kernel. If your arch is amd64, boot with the gentoo64 kernel. This means you can boot a 64-bit kernel and install a customized 64-bit user land while using the provided 32-bit user land. The livedvd-amd64-multilib-20140826 version is for x86_64 only."

Comments (none posted)

FSF adds libreCMC to its list of endorsed distributions

The Free Software Foundation has announced that libreCMC has been added to its list of free GNU/Linux distributions. LibreCMC is an embedded distribution for home wireless routers.

Full Story (comments: none)

PC-BSD 10.0.3

The PC-BSD team has announced the release of PC-BSD 10.0.3, a quarterly update to the 10.0 release. "This update includes a number of important bug-fixes, as well as newer packages and desktops. Packages such as Chromium 37.0.2062.94, Cinnamon 2.2.14, Lumina 0.6.2 and more. This release also includes a CD-sized ISO of TrueOS, for users who want to install a server without X."