On Thu, 27 Sep 2007 16:35:58 +1000 Ben Finney wrote: > Rippit the Ogg Frog <rippit@oggfrog.com> writes: [...] > > I have read many of the arguments over GPLv3, and tried to weigh > > their merits, but I'm afraid they all just make my head spin. > > > > I have always been uncomfortable with the "or, at your option, any > > later version" in GPL license notices. It is most likely that if I > > choose GPLv3, I will require that specific version and not allow any > > later version. > > The big trouble with that is the headache caused to any recipients in > the future who don't necessarily have a way to contact every > significant copyright holder in the work as it exists at that future > time. > > We're seeing that now with the mass of code under GPLv2: Works that > state "or any later version" can be freely combined with code under > GPLv3, making it far more likely the work can continue to be > maintained into the future as more desirable code is released under > GPLv3. On the other hand, if someone (like, uhm... me, for instance?) thinks that the GPLv3 has a copyleft mechanism that is too weak (that is to say: it's too permissive to be a copyleft implementation that actually works), then that person will *not* be willing to allow that his/her code is combined with GPLv3'ed code. He/she therefore will release his/her work under the GPLv2 only, and the incompatibility with GPLv3 will be considered as a feature, rather than a bug. Personally, when I want a copyleft, I want one that actually works (i.e.: GPLv2 only), not one that will allow combination with AfferoGPL (e.g.: GPLv3 only, GPLv3 or later, GPLv2 or later, ...). When I do *not* want a copyleft, I don't use overly long and complicated licenses like GPLv3, I'd rather choose a short, clear and simple non-copyleft license (e.g.: Expat). > > Works under GPLv2 that don't have that clause cannot be combined with > GPLv3 works without the long trek down a list of copyright holders, > and discussions with every one of them, which may range from "very > tedious" to "completely impossible" depending on the availability of > the copyright holders. Yes, that's the problem with a FSF that cannot be trusted anymore in keeping its promises (like the one written in section 9 of GPLv2: new versions will be "similar in spirit to the present version")... :-( [...] > Giving that permission in advance requires that you trust the FSF will > honour the spirit of the existing version of the GPL in any future > version. Exactly! That is indeed the key question: "do you trust the FSF?" > That was a commonly-held worry when they announced they were > developing version 3 of the license, but I think the process was good > at soliciting lots of feedback from stakeholders and incorporating > that into the new version. I disagree, the process was good at soliciting lots of feedback (even though with low accessibility interfaces...), but not as good at listening to that feedback. The revision process started from a first draft that was full of horribly restrictive non-free clauses and a broken copyleft and ended up with a final text that meets the DFSG (but is so close to the boundary, that sometimes I still ask myself "does it really meet the DFSG?") and a slightly less broken copyleft. A great improvement, in some sense, but I still wonder how one can trust the FSF anymore, taking into account where the GPLv3 started from (that is to say: what the FSF could include into the first draft, a license text with Monsters Inside(TM)), without mentioning the GFDL fiasco... > > Your opinion might vary, of course, and as a copyright holder it's > your decision to make for works you have monopoly in. I merely point > out that it's *far* more friendly to recipients of your work years > later if you have the "or any later version" clause. Well, a copyleft is always less recipient-friendly than a substantially do-as-you-like non-copyleft license (such as Expat, BSD, X11, ...). Consequently, no surprise that a broken copyleft turns out to be more friendly than a real copyleft! > > > Is there a web page somewhere that distills the essence of each of > > the arguments, pro and con? I've tried to read some of the threads > > and Debian legal, but I'm afraid I haven't been able to reach a > > conclusion by doing so. My analysis is just my opinion, and can be found here: http://lists.debian.org/debian-legal/2007/06/msg00267.html I am not aware of any summary of the various comments and opinions, though. [...] > > Software isn't the only thing that should be free. > > Music should be too, as it once was. > > Music in digital form, being digitially-stored information, *is* > software :-) As Eben Moglen said, it's foolish to try to distinguish > different freedoms for a bitstream depending on how those bits happen > to be interpreted at some point in time. I couldn't agree more. See my essay about the meanings of the term "software": http://frx.netsons.org/essays/softfrdm/whatissoftware.html Important disclaimers, as usual: IANAL, TINLA, IANADD, TINASOTODP. -- http://frx.netsons.org/doc/nanodocs/testing_workstation_install.html Need to read a Debian testing installation walk-through? ..................................................... Francesco Poli . GnuPG key fpr == C979 F34B 27CE 5CD8 DC12 31B5 78F4 279B DD6D FCF4
Attachment:
pgpuscasBDydR.pgp
Description: PGP signature