Re: Bug#669356: electricsheep unsuitable for Debian main?

2012-04-21 Thread Clark C. Evans
Linus,

So, it's my opinion that there are two core requirements for
free software: the license needs to be free and the whole work
must be included.  What follows is my personal opinion, and 
I'm not a lawyer, a representative of Debian Legal, or providing
any sort of legal advice.

Whole Work
--

If the software is completely usable without _requiring_ 
specific non-free parts for its operation, then you've got 
the whole work.  It is the users' right to mix or use the 
work in any way they wish, including with proprietary content 
files that may be downloaded.

On Fri, Apr 20, 2012, at 02:42 AM, Linus Lüssing wrote:
 However as far as I know the electricsheep package currently 
 heavily relies on non-free content to function properly which 
 could make it unsuitable for Debian main.

Providing the user the ability to cause the software to download
and use non-free resources on a website is quite fine, so long as 
the screen saver's software would completely work as the user
might expect without requiring those non-free resources.

If the work is effectively crippled without the connection to
their proprietary content, then they've not licensed the whole
work so as a matter of policy, I'd prefer partial-works not be
included in Debian.  In this case, it's not uncommon for someone
to fork the (incomplete) GPL'd work, remove non-free (CC licensed)
parts and contribue a working free software program.  The authors
should know this is a real possibility by using the GPL license.

  The videos downloaded and displayed by Electric Sheep are Creative
  Commons licensed (a mixture of CC-BY and CC-BY-NC).  Some jobs
  rendered by the network may be for images or animations which are not
  sheep at all, and will not appear in the screen-saver.  Some of these
  are used for commercial purposes in order to support the developers
  and servers that make the software.

I don't see the point.

In my personal opinion, if the web-browser analogy holds (that 
Electric Sheep is a browser for screen saver videos), then 
this should be unnecessary.  In this case, the user would have 
the ability to configure where it could get additional screen
saver materials at their own discretion.

Free License
-

What is meant by sheep generated by the algorithem on the 
http://electricsheep.org/reuse page?  Are these content files
that are downloaded?  If so... is there any value to the software 
besides connecting to a proprietary website?

If not, and the automatically generated sheep are part of the 
whole work that is being licensed, there is a conflict between
what this clarification page and standard GPL license they use.

If they mean to restrict the output of their program such that
it is under the by-nc, then their license should have a non-free 
term with this restriction, in which case, their software isn't 
free (or even open source).  It definately isn't GPL licensed.

Otherwise, if they intend to license their program under the 
terms of the GPL, then the output of this program is unencumbered 
and they should remove their clarification comment about sheep 
generated by the algorithem since it is incorrect.

I hope this helps.  This stuff can be complex so I'm sure others
may have a different take on it.

Best,

Clark


--
To UNSUBSCRIBE, email to debian-legal-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: 
http://lists.debian.org/1335024173.24581.140661065450701.55f8e...@webmail.messagingengine.com



Re: No source code for wesnoth-music

2012-04-21 Thread Francesco Poli
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


pgpzrFAxkLAqw.pgp
Description: PGP signature


Re: No source code for wesnoth-music

2012-04-21 Thread Francesco Poli
On Mon, 16 Apr 2012 20:56:55 -0500 Guilherme de Siqueira Pastore wrote:

 On Mon, Apr 16, 2012 at 07:57:39PM +0200, Francesco Poli wrote:
  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.
 
 Law is not a computer algorithm, and the intention of the licensor must be
 taken into consideration.

I agree, but the intention is in general not overly clear, especially
when a licensor releases a work under the GPL and, at the same time,
does not make source available.
Self-contradicting actions leave the beholders in doubt about the real
intentions.

Please also take into account that, in some cases, the authors of some
wesnoth-music files explicitly *refused* to release source (== their
preferred form for making modifications), and proposed, as a solution
to re-license under the terms of a license which does *not* have any
source-availability requirement. So, at least in those cases, their
intentions to *not* make source available is pretty clear.
See the previously cited forum threads...

 If licensing the software under the GPL indicates
 a clear intent to allow redistribution, provided that the same freedom you
 were granted by the copyright holder is passed on to those who get the
 software from you, the fact that a small portion of the overall software
 package has no source code available (even to the copyright holder) should not
 bar such redistribution.

I don't know how it would end up in court, in case someone [*] sues,
but I think that, in order to be on the safe side, we should assume a
strict interpretation of the license and consider those files as
legally undistributable.

[*] by someone, I mean the copyright holder(s) of other parts
of the package, not the authors of the unreleased source files

 
 Under certain civil law jurisdictions, it would also be technically accurate
 to state that conditions which cannot possibly be met are void, although they
 do not void the rest of the contract of which they are part.
 
 Again, making it clearer does not hurt and certainly avoids potential
 future discussion, but I do not agree there is any risk of liability by not
 distributing something the copyright holder (inadvertently) required you to
 redistribute, without ever distributing it to you.

Would you reason the same way, if we were speaking about a GPL-licensed
executable program, distributed in binary form, while keeping its
source code secret?

 
 
  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.
 
 I am not discussing whether the music and image files in the Debian archive
 should comply with the DFSG requirements or not - I have been a strong
 advocate thereof since the early days of my involvement with Debian, back in
 2002. What I am arguing is whether wesnoth music and image files would be
 considered non-free, given the circumstances.

For the cases where it is clear that the source (== preferred form for
making modifications) is not what is being distributed and the original
authors explicitly stated that they do *not* want to release it to the
public, ..., well, how shall we call these files? differently free?!?

[...]
 However, I do not think that attaching the non-free label to an effectively
 free piece of software is consistent with the other goals we have agreed on
 as a community, especially if there is no unwillingness on the upstream
 author's part to fix a license problem, or any attempt to circumvent the
 requirements of the DFSG and deprive our users of their freedoms, but rather
 an unfortunate lack of caution with the handling of file formats.

In some cases there appears to be an unwillingness to fix this issue,
as the authors actively refused to release source.

In other cases, the form originally used to generate the distributed
file seems to no longer exist. If this is really true, then the source
for such a file is clearly one of the actually existing ones: possibly
the one which is being distributed; if this is the case, no problem
whatsoever.

 
 These are considerations that apply to the specific circumstances of the given
 case. I am not trying to revert the project consensus (and that might be
 related to the fact that I agree with it!), just trying to be reasonable.

I am also trying to be reasonable: I think that we should not set
precedents that qualify secret-source files as acceptable in Debian
main, just because they are not programs...

 
 Quoting from another GR proposition (that did not pass), source code can be
 understood as the form that the copyright holder or upstream developer would