| |||||||||||||
|
| |||||||||||||
A new GPLv3 timetableA new GPLv3 timetablePosted Mar 27, 2007 14:24 UTC (Tue) by flewellyn (subscriber, #5047)In reply to: A new GPLv3 timetable by rknop Parent article: A new GPLv3 timetable
They're leery of it because of the "front-cover text", "back-cover text", and "invariant section" clauses.
I can kinda see both sides of the issue here. On the one hand, FSF put those in because they want the ability to distribute, say, the GNU Manifesto, or the book's license, with a GFDL-covered book. Obviously, they don't want those to be changeable, so they would mark them as "invariant". Similar with the front and back-cover texts, which are apparently meant for the title and author or something similar.
On the other hand, while FSF may well restrict themselves to only marking those particular things invariant, the license is not sufficiently clear on what can and can't be so marked. There's concern that GFDL manuals or books could have important parts marked "invariant", when those parts should be changeable.
I think FSF made a bad decision when they put in the front-cover, back-cover, and invariant section clauses. I understand their reasoning, but I think those clauses are the wrong way to handle the problem. What I would do is simply remove those clauses, and then distribute parts of the book that are meant to be "invariant" under a different license altogether, and prominently mark those parts of the book (which are separate anyway) as being "not under the GFDL". Generally the FSF puts the license and the manifesto in appendices, in completely separate parts of the book from the actual manual text.
I see nothing wrong with doing so, declaring that those parts are "copyrighted but allowing unlimited distribution" (which text is already found in the licenses and the Manifesto), provided that the text of these "invariant" sections is kept strictly separate from the rest of the book. Conceivably this could be justified as "mere aggregation", in the GPL's sense.
(Log in to post comments)
A new GPLv3 timetable Posted Mar 27, 2007 16:09 UTC (Tue) by error27 (subscriber, #8346) [Link] Invariant sections aren't bad in themselves. It's just that the language of the GFDL is bad and imprecise. You could have a GFDL doc that was entirely invariant.
A new GPLv3 timetable Posted Mar 27, 2007 18:42 UTC (Tue) by k8to (subscriber, #15413) [Link] Personally I do not see why it is obviously desirable to mark the GNU Manifesto unchangable. I can copy the text today and deface it without agreeing to any portion of the GFDL under fair use for parody or whatnot. Save defacement, what changes are going to occur to the text? Removal is about the only other likely (in fact MORE likely) change. The only reason that it makes sense to make the GNU manifesto invariant and thus unremovable, is if the FSF believes that the manifesto is text that users do not want to hear. This very act creates an atmosphere of assumed desired rejection and I think is contrary to their goals.
As for invariant covers, this essentially attempts to prevent the reuse of GNU documentation in other documents, mostly likely with an eye to commercial physical-book publications. Again, I cannot see that it is desirable to discourage reuse of the GNU documentation, as it would cause the documentation to be more broadly disseminated which would encourage adoption of the GNU tools. The downside is a mere protection of an established GNU revenue stream: the printed copies of GNU manuals. Personally I am doubtful that many of the GNU manuals are purchased solely for utility reasons in a manner that would be so easily interrupted. But if this is the case, cannot the FSF seek to diversify their revenue sources, so that the documentation they create can eventually be as Free as the software it is designed to accompany?
A new GPLv3 timetable Posted Mar 28, 2007 2:01 UTC (Wed) by njs (subscriber, #40338) [Link] >They're leery of it because of the "front-cover text", "back-cover text", and "invariant section" clauses.
That's not entirely complete... a good bit of reading for those that are interested is the summary prepared by debian-legal back during the original debate:
Invariant sections are the most gratuitous way that the GFDL is DFSG-non-free, but a lot of reasonable people believe that it still isn't DFSG-free even when there are no invariant sections. This isn't the official position of the Debian project -- they decided that only invariant sections are a problem -- but that's not because the debian-legal denizens reached this as consensus, but rather that the project as a whole held a popular vote and the majority declared that it was free by fiat. I'm not entirely impressed by this. I find the arguments on that web page compelling, and am anyway more inclined to trust the experts than the popular opinion; if I were a debian developer I would have voted for the GFDL being simply non-free.
Also, even if you do consider the GFDL free, there are a giant list of ways that the GFDL is just obnoxious in practice. (One major one is that it is itself GPL incompatible.) That doesn't matter from a distribution's point of view, but it's important to keep in mind if you are producing a new work and thus have to choose a license. These days I just use GPL for documentation; at least that way I know I can copy and paste between it and the code... (I've not been following the Creative Commons license revisions, does anyone know how acceptable they are these days?) Personally I would strongly recommend against anyone choosing to use the GFDL. At the very least, you should think about it very carefully; I have a huge amount of respect and trust for the FSF, but I think they really screwed this one up.
A similar and thoughtful discussion of these issues that influenced me: http://zwol.livejournal.com/13793.html
A new GPLv3 timetable Posted Mar 28, 2007 21:28 UTC (Wed) by flewellyn (subscriber, #5047) [Link] Hm, I wasn't aware of the issues with the DRM section, or with the transparent and opaque sections. I only ever heard about the objections to invariant sections, front covers, and back covers.
That said, I think the new draft SFDL is a better idea, since it doesn't include invariants, front covers, or back covers. I still think FSF could distribute the Manifesto and the license texts as separate, aggregated works.
|
|
||||||||||||