On Wed, Mar 05, 2003 at 08:06:05PM +0100, Henning Makholm wrote: > Scripsit Steve Langasek <vorlon@netexpress.net> > > Does anyone here hold the position that requiring the copyright notice on > > the front page would not be DFSG-free, if that's a valid interpretation > > of the GPL? > I have a hard time deciding what to think here. On one hand, it is > not, in the present case, an orneous requirement. On the other, I fear > a slippery slope leading to output restrictions and all sorts of bad > things. > Perhaps a reasonable place to draw the line would be to say that we > don't find it DFSG-unfree in itself to require a copyright notice (or > other stuff as detailed in GPL 2(c) on the front page, if the software > appears *to the remote end user* as a coherent entity that he is > having a stateful interaction with. This would be independent of how > that appearance is technically achieved. However, for some technical > architectures, the means by which the copyright notice is required in > the license may still be unfree in themselves - for example, if they > are phrased as an output or use restriction. I'm not sure you've answered the question I meant to ask. Let me try to rephrase: if debian-legal finds that such a requirement from upstream is a legitimate clarification of the GPL (rather than an additional restriction imposed on top of the GPL), do you think it's appropriate for debian-legal to reject a piece of GPL software whose author imposes this restriction, given that the GPL is explicitly grandfathered into the DFSG? Whether this particular requirement is a clarification or an additional restriction is a separate question, I think (my third one, repeated below). I believe that if we get to the point where debian-legal agrees that an interpretation of the GPL is valid, but does not agree that this same interpretation is sufficiently free to be allowed in Debian, it's time to revise the DFSG by means of a GR. > > Does anyone believe that this interpretation is sufficiently wrong-headed > > that it should not be considered valid, in spite of statements from the > > copyright holder or a court ruling? > For debian-legal's purposes, I suppose that the constructive position > would be to *assume* that the upstream author is right when he claims > that the GPL includes restrictions that we usually don't see the GPL > as including. I think it is always appropriate to assume the license on a piece of software is exactly what the copyright holder states that it is; if nothing else, this avoids unnecessary lawsuits. I don't think it's appropriate to always accept the copyright holder's statements as an *interpretation of the GPL*. I think we need to decide independently what is a valid interpretation of the GPL, and what is an additional restriction imposed by the copyright holder -- specifically because of the GPL grandfathering. > If the restrictions claimed by the upstream author are > unacceptable, we should treat the software as DSFG-nonfree even though > the restrictions may not actually be there, seen from an objective > legal viewpoint. This, in spite of the DFSG's GPL grandfathering? > Things only begin to get muddy in the hypothetical case that the > project has been forked or inherited by a new upstream author, and the > new upstream author interprets the GPL to grant rights that the old > one explicitly denied. *Then* we'll nedd to figure out which of the > authors we believe. I don't think this is muddy at all: the effective license on the code is that given by the original copyright holder. A new maintainer can't grant us more rights to code he doesn't own. -- Steve Langasek postmodern programmer
Attachment:
pgp5Rc5XMHV36.pgp
Description: PGP signature