On Tue, 17 Apr 2012 13:52:38 -0500 Gunnar Wolf wrote: > Francesco Poli dijo [Mon, Apr 16, 2012 at 07:57:39PM +0200]: > > > > but anybody else breaches the GPL 2 or 3 by not offering source code. > > > > > > ... but I tend to disagree at this point. Despite the possibility of > > > considering these file types as source code for the relevant purposes under > > > the circumstances, I am not sure we can talk about license violation from a > > > legal standpoint if the infringed requirement is that of redistribution of > > > something the redistributor never received (and, in this case, something > > > even the copyright holder could not provide, because it does not exist). This > > > should, at the very least, constitute grounds for exemption of liability. > > > > I am not convinced: if someone releases a work under the GPL without > > making the corresponding source available, nobody else really has the > > true permission to redistribute, as the license requires > > re-distributors to make source available, but they cannot, since they > > do not have it in the first place. > > Well, where does the source code requirement of the GPL come from? I'd > say, based on the FSF's famous four liberties, that quite probably > from the conjunction of "freedom to learn" and "freedom to > modify". For a C program (that is, FSF's initial area of interest) to > be learnable and modifiable, you clearly need the source code. Yes, I would say. > > Now, music is not learnt directly through its sources (i.e. a MIDI > file and the used samples). Not *only* through its source (== preferred form for making modifications), but *also* through other forms. But the same is true for a C program: you can learn a lot about it, by using its executable binary, studying its documentation, testing it under odd conditions, analyzing how it interacts with other programs and/or the network (for instance, think about the use of network "sniffers" to understand what a network application does), de-compiling it, disassembling it, and so forth... > And it can be meaningfully modified. The same holds for a C program: you can de-compile/modify/re-compile it, you can disassemble/modify/re-assemble it, edit it with a hex-editor, and so forth. > So, > yes, we talk about a field where "modifiability" has many gradients. I > agree that the OGG files are quite possibly not the "prefered form of > modification" (specially for any synthetized music - I don't know > Wesnoth or its music). In some cases a music file in OGG Vorbis format may be its own source. In other cases it is not. It really depends on the specifics of the situation under consideration. Please note that the same holds for a program: sometimes C code is source, sometimes it's automatically generated from code written in another language (for example a grammar description language), which is the real source. > > My take on this would be, the Debian maintainer responsible for a > given program should ask its upstream for something that qualifies as > source, but if upstream refuses (or just says it does no longer exist > — Effectively the same), continue to distribute what we have. I disagree. If a form really no longer exists, not even in the hands of the original author (or copyright holder) of the work, then such form cannot qualify as source (== preferred form for making modifications to the work). Whenever this is the case, we must determine which is the preferred form *among the existing ones*. If instead a form still exists, is the form that the author prefers for making modifications to the work, but is kept secret and unreleased (by either refusing to disclose it or by falsely claiming that it no longer exist), then we are dealing with a secret-source work, which does not comply with DFSG#2. > > > There was a GR in 2004 to clarify the social contract, in order to make > > it clear that the DFSG apply to all works (in Debian main), not just > > executable programs. > > Hence I think the agreed upon interpretation is that music and images > > must include source code. > > Right. The problem is the definition of source code. Please let's not restart the "what is source?" debate! I may consider it as expected on the cc-licenses mailing list, but not here on debian-legal! -- http://www.inventati.org/frx/frx-gpg-key-transition-2010.txt New GnuPG key, see the transition document! ..................................................... Francesco Poli . GnuPG key fpr == CA01 1147 9CD2 EFDF FB82 3925 3E1C 27E1 1F69 BFFE
Attachment:
pgpe4KVaEGE00.pgp
Description: PGP signature