July 20, 2015
Additional Participants: Chris Mason, David Howells, Greg KH, Guenter Roeck, John W. Linville, Steven Rostedt, Tim Bird, and Tony Luck.
People tagged: Bradley Kuhn, Karen Sandler, and an LF TAB representative.
David Woodhouse “slightly loathly” brought up the topic of GPL enforcement, and with it the meaning of the GPL. David said that some parties are portrayed as excessively and inappropriately enforcing GPL, while others are portrayed as being rabidly against any enforcement action that might differentiate GPL from the BSD license. David is concerned that the histrionics stemming from this conflict will prove damaging, and suggests that a coherent discussion might result in a useful “feeling in the room” regarding appropriate compliance with GPL.
Steven Rostedt argued that this topic would require that a lawyer be present, in order to provide some reality as to what the law actually entails. Steven also noted the additional complexities of different laws in different jurisdictions. David Woodhouse countered that lawyers tend to provide the opinions that they are paid to provide, and that that there is no “right” answer until the issue is seen in court, and Greg KH seconded this. [ Ed. note: This is especially true in countries having “case law” such that court cases have legal force over and above the wording of the actual laws themselves. Both the UK (where David lives) and the USA (where your editor lives) are case-law contries. ] David further argued that gaining a sense of the community's desires would be far more productive than would a purely legal discussion. David suggested that Bradley Kuhn or Karen Sandler of the Software Freedom Conservancy and a representative from the The Linux Foundation Technical Advisory Board be present to give their views on GPL enforcement actions. Greg KH pointed out that these actions are being driven by kernel developers, and suggested that having those kernel developers present would be more productive. David countered that the Software Freedom Conservancy is actually doing the work, and would thus be in a better position to say what is actually happening, a notion that John W. Linville seconded, but Tim Bird stated that the resulting discussion would likely be one-sided, as the companies on the receiving end of GPL enforcement actions would be unlikely to comment. Tim nevertheless agreed that it would be useful to hear Software Freedom Conservancy's perspective. Greg GH argued that companies on the receiving end of GPL enforcement actions would be unable to comment, rather than merely unlikely to do so, and asked what specific issues Tim was interested in hearing about from Software Freedom Conservancy. Tim replied that he is interested in the scope of remedies requested by the Software Freedom Conservancy, specifically, whether if a violation by a specific group in a company for a specific project would likely result in some sort of review-and-approve process for the rest of the company and for other projects. Tim added that it was not clear whether or not this discussion would be useful. Chris Mason suggested that a survey might work better than would an open discussion, which he suspects would be dominated by those with extreme positions. Chris also stated that he would prefer a fact-based session covering current enforcement actions to an open discussion on GPL.
Tony Luck noted that most Linux-kernel commits come from people paid to hack the Linux kernel, and asked if their employers should have a say, noting that the employers and employees might have different views on this topic. Chris Mason argued that the employers had had a say all along, and stated that the Linux-kernel community's stability had in fact benefited from their say. Guenter Roeck suggested that the Linux Foundation attorneys brief their member companies about the meeting's outcome.
David Howells pointed out a related topic, namely the fact that many important exported symbols are being marked GPL-only, which affects not just closed-source kernel modules, but also non-GPL open-source projects such as OpenAFS. David Woodhouse admitted that this was perhaps related, and asked if David Howells hadn't finished obsoleting the kernel parts of OpenAFS. :)