(no subject)
Mar. 12th, 2006 08:41 amI have severely mixed feelings about the result of the Debian vote on the freeness of the GFDL.
Short summary for those unfamiliar with this issue: The GFDL is the Free Software Foundation's recommended license for documentation, and is used by a number of other organizations, e.g. Wikipedia. It has a number of provisions which are, let us just say, controversial. Debian, well known for its uncompromising adherence to the principle of distributing only free-as-in-speech bits, has been engaged in a years-long internal dispute over whether bits under the GFDL can be included in their distribution, which has culminated in the vote whose results are linked from the first paragraph. It is useful to read the Debian Free Software Guidelines, which set the field of debate: all talk of "is it free enough?" is in reference to these guidelines. It's also important to know that the license was presumed innocent while the debate was going on, so there is currently quite a lot of GFDL-licensed documentation in Debian.
I'm not a Debian developer. I am, however, the author of a fairly large hunk of text (the GNU CPP manual) which is distributed under the GFDL — not because I wanted it to be, but because the FSF insists you use that license for GNU project documentation (which is their right to do, since it's their project). At the time I wrote the manual, I hadn't considered the license carefully. Since then I've been convinced by the arguments of the "this license is completely unacceptable" faction within Debian, and if I could retract the contribution of my manual without causing enormous collateral damage, I would. All this is to say that I've been following the debate pretty closely.
The vote result is a compromise: only GFDL works with "unmodifiable sections" (the ballot's term; this corresponds to "Invariant Sections" and/or "Cover Texts" in the GFDL's terminology) are to be thrown out of Debian. Now, on the one hand, unmodifiable sections really are the worst provision of the GFDL; all the other problems are at least arguably just poor wording and/or failure to consider all the consequences, and perhaps should not be held against the license. Or, to put it another way, authors using the GFDL without specifying unmodifiable sections are probably doing so in good faith that it's a free license; like me, they probably didn't think about it very much and assumed that the FSF wouldn't promulgate an unfree license. So it may be reasonable to leave that documentation in Debian.
On the other hand, most of the documentation in Debian under the GFDL comes from the FSF, and the FSF uses unmodifiable sections in everything it releases under that license. Thus, the practical impact on users is almost as severe as if all GFDL documentation were removed from Debian. Furthermore, at several points during the aforementioned internal debate, people tried to involve the FSF in the discussion; it rapidly became clear that the FSF would not hear any objections to the license. [Since then, several people have claimed to be continuing negotiations with them in private with some success, but there has been no visible effect, and personally I do not believe there has been any real progress.] Partially this was because Debian was the first group to raise objections, and the FSF didn't appear to believe they were serious. Since then I have been hoping that Debian would indeed eject all GFDLed bits from the distribution, as that would at least force the FSF to admit that Debian did have serious problems with the license. By ejecting only the documents with unmodifiable sections, Debian weakens its position considerably; it can appear that it's the unmodifiable sections that are the problem, not the license itself.
Finally, this result is a partial victory for the faction within Debian that argued for the as-is acceptability of the GFDL. That is a legitimate debate position, but that faction has not debated in good faith. They ignore or deliberately misread rebuttals of their position, they engage in ad hominem attacks, and they repeat their claims over and over until their opponents give up responding. They should not be rewarded for these tactics. (To be fair, some of the people in the other faction have done these things too, but everyone on the it's-fine-as-is side has done these things, especially the last.)
Short summary for those unfamiliar with this issue: The GFDL is the Free Software Foundation's recommended license for documentation, and is used by a number of other organizations, e.g. Wikipedia. It has a number of provisions which are, let us just say, controversial. Debian, well known for its uncompromising adherence to the principle of distributing only free-as-in-speech bits, has been engaged in a years-long internal dispute over whether bits under the GFDL can be included in their distribution, which has culminated in the vote whose results are linked from the first paragraph. It is useful to read the Debian Free Software Guidelines, which set the field of debate: all talk of "is it free enough?" is in reference to these guidelines. It's also important to know that the license was presumed innocent while the debate was going on, so there is currently quite a lot of GFDL-licensed documentation in Debian.
I'm not a Debian developer. I am, however, the author of a fairly large hunk of text (the GNU CPP manual) which is distributed under the GFDL — not because I wanted it to be, but because the FSF insists you use that license for GNU project documentation (which is their right to do, since it's their project). At the time I wrote the manual, I hadn't considered the license carefully. Since then I've been convinced by the arguments of the "this license is completely unacceptable" faction within Debian, and if I could retract the contribution of my manual without causing enormous collateral damage, I would. All this is to say that I've been following the debate pretty closely.
The vote result is a compromise: only GFDL works with "unmodifiable sections" (the ballot's term; this corresponds to "Invariant Sections" and/or "Cover Texts" in the GFDL's terminology) are to be thrown out of Debian. Now, on the one hand, unmodifiable sections really are the worst provision of the GFDL; all the other problems are at least arguably just poor wording and/or failure to consider all the consequences, and perhaps should not be held against the license. Or, to put it another way, authors using the GFDL without specifying unmodifiable sections are probably doing so in good faith that it's a free license; like me, they probably didn't think about it very much and assumed that the FSF wouldn't promulgate an unfree license. So it may be reasonable to leave that documentation in Debian.
On the other hand, most of the documentation in Debian under the GFDL comes from the FSF, and the FSF uses unmodifiable sections in everything it releases under that license. Thus, the practical impact on users is almost as severe as if all GFDL documentation were removed from Debian. Furthermore, at several points during the aforementioned internal debate, people tried to involve the FSF in the discussion; it rapidly became clear that the FSF would not hear any objections to the license. [Since then, several people have claimed to be continuing negotiations with them in private with some success, but there has been no visible effect, and personally I do not believe there has been any real progress.] Partially this was because Debian was the first group to raise objections, and the FSF didn't appear to believe they were serious. Since then I have been hoping that Debian would indeed eject all GFDLed bits from the distribution, as that would at least force the FSF to admit that Debian did have serious problems with the license. By ejecting only the documents with unmodifiable sections, Debian weakens its position considerably; it can appear that it's the unmodifiable sections that are the problem, not the license itself.
Finally, this result is a partial victory for the faction within Debian that argued for the as-is acceptability of the GFDL. That is a legitimate debate position, but that faction has not debated in good faith. They ignore or deliberately misread rebuttals of their position, they engage in ad hominem attacks, and they repeat their claims over and over until their opponents give up responding. They should not be rewarded for these tactics. (To be fair, some of the people in the other faction have done these things too, but everyone on the it's-fine-as-is side has done these things, especially the last.)
![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}