public inbox for gcc@gcc.gnu.org
 help / color / mirror / Atom feed
* Re: [3.2/3.3/HEAD] Make all the manuals unambiguously DFSG free
       [not found] <20030123150602.A15387@synopsys.comþ2>
@ 2003-01-24 11:53 ` Kai Henningsen
  0 siblings, 0 replies; 9+ messages in thread
From: Kai Henningsen @ 2003-01-24 11:53 UTC (permalink / raw)
  To: gcc

jbuck@synopsys.com (Joe Buck)  wrote on 23.01.03 in <20030123150602.A15387@synopsys.com·2>:

> In any case, it appears that the FSF and the Debian folks may find
> themselves at an impasse: the FSF wants to be able to spread its message
> ("propaganda" for those who don't like it) by including it in manuals, and
> Debian seems to be trying to extend the DFSG to non-software and wants to
> bar attempts to do this.

Just a small clarification of the ebian side: there is no expansion going  
on here. The DFSG were *always* meant to apply to everything in Debian; it  
is only recently that people noticed that much documentation has  
conflicting licences. When the DFSG were created, most everyone probably  
assumed that documentation to a program was usually under the exact same  
terms as the source, and that this was certainly true for FSF packages - I  
know I was. Come to think of it, I'm rather certain the DFSG are quite a  
bit older than the GFDL ...

Then people became astonished and upset at realizing that the software  
package that inspired the creation of the GPL (which is probably the main,  
at least motivational, influence on the DFSG), Emacs, is one of the worst  
offenders with its must-keep-propaganda clause. While I've given up on  
trying to have any kind of position in the debate of what to do about the  
DFSG-conflict problem, I will say that I am also one of those who feel  
this stinks of hypocrisy, and who completely fail to understand why there  
should be a material licensing difference between a program and its  
documentation.

In any case, this is not the place to discuss Emacs, so the above is only  
meant to shed some light on the background of this debate.

MfG Kai

^ permalink raw reply	[flat|nested] 9+ messages in thread

* Re: [3.2/3.3/HEAD] Make all the manuals unambiguously DFSG free
  2003-01-24  1:52 ` Joseph S. Myers
@ 2003-01-24  5:32   ` Joe Buck
  0 siblings, 0 replies; 9+ messages in thread
From: Joe Buck @ 2003-01-24  5:32 UTC (permalink / raw)
  To: Joseph S. Myers; +Cc: Richard Kenner, gcc

I skipped the part about the Ada manual and the GFDL, that's already
being fixed, based on communication between the FSF and the ACT folks.

On Thu, Jan 23, 2003 at 11:40:29PM +0000, Joseph S. Myers wrote:
> Must explicit FSF approval be given to add copyright and licence notices
> to files missing them?

The FSF's "copyright clerk" regularly conducts audits of copyright
statements and occasionally points out missing copyrights to the SC list
(and also tries to flag code that includes other code with an apparently
contradictory license).  We then fix them.

> (For example, the FSF says that any file than ten
> lines long, including such files as ChangeLogs, should have such notices -
> a simple permissive licence being appropriate for rough documentation such
> as ChangeLogs.)  Must it be given to add the libgcc exception to a libgcc
> file wrongly lacking it?  Must it be given to copy code from a non-libgcc
> file into libgcc?

For existing libraries, the FSF has already agreed to the general terms,
e.g. that libgcc has a certain license.  In cases like that, we already
have permission to fix files that lack the needed exception, because we're
just making things match what the FSF has already agreed to.

The point is that the FSF makes the macro-decision, and doesn't need to be
bothered with all the micro-decisions that follow from that (e.g. choosing
the license of 50 files that all form one library).

However, we need to be careful: if we want to take code out of library A
(with one license) and put it in library B (with another license), then
usually the SC will have to talk RMS into doing it, and sometimes this
can take quite a while.  Our latest issue was getting RMS to let us
relicense the C++ demangling function, since it had to make such a move.

Also, things that are clear to us (that language support libraries have to
have a less restrictive license than the GPL) may not be clear to RMS.
We had a knock-down drag-out fight for about a year to clean up the issues
surrounding the Java awt licensing situation.


^ permalink raw reply	[flat|nested] 9+ messages in thread

* Re: [3.2/3.3/HEAD] Make all the manuals unambiguously DFSG free
  2003-01-24  1:58   ` Zack Weinberg
@ 2003-01-24  5:24     ` Joseph S. Myers
  0 siblings, 0 replies; 9+ messages in thread
From: Joseph S. Myers @ 2003-01-24  5:24 UTC (permalink / raw)
  To: Zack Weinberg; +Cc: gcc

On Thu, 23 Jan 2003, Zack Weinberg wrote:

> The controversy here is external to the GCC Project.  The argument is
> strictly speaking between thee Debian project and the drafters of the
> FDL.  My belief is that the GCC project should avoid getting into the
> argument, and my patch was intended to achieve such avoidance - no
> more.

Getting into such argument is avoided by closing any FSF GCC bug reports
about the standard GNU applications of the FDL with a reference to FSF
policies as described in "GNU Sample Texts" in the Texinfo manual [0] and
the FSF's instructions on using the FDL
<http://www.gnu.org/licenses/fdl-howto.html>, further questions being
directed to RMS (though he doesn't have a GNATS account so you can't
assign PRs to him).

The FSF's licensing policies are up to the FSF, and Debian's are up to
Debian.  If neither will move then the FSF can always change its preferred
GNU/Linux distribution to one including the GCC manuals if it wants to and
Debian removes the manuals.

[0] At present the manuals use an older version of these with the
Front-Cover Text "A GNU Manual" being given out of line, "the Front-Cover
texts being (a) (see below)"; the current version gives the Front-Cover
Text inline and only the Back-Cover Text out of line.

-- 
Joseph S. Myers
jsm28@cam.ac.uk

^ permalink raw reply	[flat|nested] 9+ messages in thread

* Re: [3.2/3.3/HEAD] Make all the manuals unambiguously DFSG free
  2003-01-24  2:01 Richard Kenner
@ 2003-01-24  4:51 ` Joe Buck
  0 siblings, 0 replies; 9+ messages in thread
From: Joe Buck @ 2003-01-24  4:51 UTC (permalink / raw)
  To: Richard Kenner; +Cc: jsm28, gcc

On Thu, Jan 23, 2003 at 08:01:21PM -0500, Richard Kenner wrote:
> My own guesses as to answers to your questions:

Pretty much for any of these questions, 
>     Must explicit FSF approval also be given for fixing _defective_
>     licence applications?  
> 
> It depends on how obvious it is that there's a "defect".

.. especially if the proposed fix is to
>     Some of the Ada manuals do not apply the GFDL in the proper GNU way;
> 
> The Ada manuals are somewhat of a special case because they are basically
> in the process of being converted to "the proper GNU way".
> 
>     When new files (covered by an assignment) are added, may they in all
>     cases be added under the standard FSF terms (as described at
>     <http://www.gnu.org/prep/maintain_9.html>) without explicit approval?
>     What about other cases - may a file that will form part of libgcc be
>     added with the libgcc exception without explicit approval?  If an
>     author wishes to use the GFDL for a new manual without Cover Texts,
>     may they? 
> 
> I'd argue that for new files the author specifies the copyright terms
> and those would survive the assignment to the FSF so that the answer to
> all of those would be "yes" unless there was some clear conflict with
> FSF practice.
> 
>     Must explicit FSF approval be given to add copyright and licence
>     notices to files missing them? 
> 
> I'd say "no" because from a legal point of view, those files already have the
> standard terms applied anyway even without those notices.
> 
>     Must it be given to add the libgcc exception to a libgcc file wrongly
>     lacking it?  Must it be given to copy code from a non-libgcc file into
>     libgcc?
> 
> Technically, "yes", but the "de mimimus" and "clear error" doctrines are
> relevant here and may mean the answer is "no" in specific case.

-- 
Q. What's more of a headache than a bug in a compiler.
A. Bugs in six compilers.  -- Mark Johnson

^ permalink raw reply	[flat|nested] 9+ messages in thread

* Re:  [3.2/3.3/HEAD] Make all the manuals unambiguously DFSG free
@ 2003-01-24  2:01 Richard Kenner
  2003-01-24  4:51 ` Joe Buck
  0 siblings, 1 reply; 9+ messages in thread
From: Richard Kenner @ 2003-01-24  2:01 UTC (permalink / raw)
  To: jsm28; +Cc: gcc

My own guesses as to answers to your questions:

    Must explicit FSF approval also be given for fixing _defective_
    licence applications?  

It depends on how obvious it is that there's a "defect".

    Some of the Ada manuals do not apply the GFDL in the proper GNU way;

The Ada manuals are somewhat of a special case because they are basically
in the process of being converted to "the proper GNU way".

    When new files (covered by an assignment) are added, may they in all
    cases be added under the standard FSF terms (as described at
    <http://www.gnu.org/prep/maintain_9.html>) without explicit approval?
    What about other cases - may a file that will form part of libgcc be
    added with the libgcc exception without explicit approval?  If an
    author wishes to use the GFDL for a new manual without Cover Texts,
    may they? 

I'd argue that for new files the author specifies the copyright terms
and those would survive the assignment to the FSF so that the answer to
all of those would be "yes" unless there was some clear conflict with
FSF practice.

    Must explicit FSF approval be given to add copyright and licence
    notices to files missing them? 

I'd say "no" because from a legal point of view, those files already have the
standard terms applied anyway even without those notices.

    Must it be given to add the libgcc exception to a libgcc file wrongly
    lacking it?  Must it be given to copy code from a non-libgcc file into
    libgcc?

Technically, "yes", but the "de mimimus" and "clear error" doctrines are
relevant here and may mean the answer is "no" in specific case.

^ permalink raw reply	[flat|nested] 9+ messages in thread

* Re: [3.2/3.3/HEAD] Make all the manuals unambiguously DFSG free
  2003-01-24  1:42 ` Joe Buck
@ 2003-01-24  1:58   ` Zack Weinberg
  2003-01-24  5:24     ` Joseph S. Myers
  0 siblings, 1 reply; 9+ messages in thread
From: Zack Weinberg @ 2003-01-24  1:58 UTC (permalink / raw)
  To: Joe Buck; +Cc: Richard Kenner, gcc-patches, gcc

Joe Buck <jbuck@synopsys.com> writes:

> Zack, try to watch your tone.  Language like "I am not interested in
> rehashing the debate", or "assertions that there is nothing wrong with the
> status quo will be taken as evidence in *favor*", come across as arrogant,
> and this is not the first time that you have used such an approach.
...
> Under the GCC Project's rules, demonstration of the existence of
> controversy on a "political" matter isn't an argument in favor of a
> change; just the opposite: it means that the SC must vote on the
> issue, and it takes a 3/4 vote to make a change.

I apologize for my tone.  My grouchiness level is elevated at the
moment for reasons totally unrelated to the subject of discussion, and
this bled into the message.  What I was trying to get across was
simply that I don't see a need to reprise the debate over whether the
FDL meets the DFSG here; it's been done to death elsewhere.

The controversy here is external to the GCC Project.  The argument is
strictly speaking between thee Debian project and the drafters of the
FDL.  My belief is that the GCC project should avoid getting into the
argument, and my patch was intended to achieve such avoidance - no
more.

> Nevertheless, if you can talk RMS into a resolution that both he and
> the Debian folks can live with, more power to you.  Come back when
> you've convinced him; I hope you're a patient man.

I have contacted him, but I don't plan to try very hard to convince
him.  If we have some assertion on his part that we can point at if
and when Debian decides that our manuals definitely don't fit their
guidelines, that will hopefully serve my goal (staying out of the
argument) well enough; and if it doesn't, we can worry about it then.

zw

^ permalink raw reply	[flat|nested] 9+ messages in thread

* Re:  [3.2/3.3/HEAD] Make all the manuals unambiguously DFSG free
  2003-01-24  1:27 Richard Kenner
  2003-01-24  1:42 ` Joe Buck
@ 2003-01-24  1:52 ` Joseph S. Myers
  2003-01-24  5:32   ` Joe Buck
  1 sibling, 1 reply; 9+ messages in thread
From: Joseph S. Myers @ 2003-01-24  1:52 UTC (permalink / raw)
  To: Richard Kenner; +Cc: gcc

On Thu, 23 Jan 2003, Richard Kenner wrote:

> You cannot make that change because you are changing the terms of a copyright
> held by the FSF.  Only the FSF can legally make such a change.

Must explicit FSF approval also be given for fixing _defective_ licence
applications?  Some of the Ada manuals do not apply the GFDL in the proper 
GNU way; rather than using the standard GNU texts, they specify the manual 
title as a Front-Cover Text (when the GFDL requires the title on the front 
cover anyway) and the GFDL as an Invariant Section (when it is invariant 
anyway and does not need to be explicitly listed as such).

When new files (covered by an assignment) are added, may they in all cases
be added under the standard FSF terms (as described at
<http://www.gnu.org/prep/maintain_9.html>) without explicit approval?  
What about other cases - may a file that will form part of libgcc be added
with the libgcc exception without explicit approval?  If an author wishes
to use the GFDL for a new manual without Cover Texts, may they?  What
about the old permissive licence for manuals (still used for
cppinternals.texi)?

Must explicit FSF approval be given to add copyright and licence notices
to files missing them?  (For example, the FSF says that any file than ten
lines long, including such files as ChangeLogs, should have such notices -
a simple permissive licence being appropriate for rough documentation such
as ChangeLogs.)  Must it be given to add the libgcc exception to a libgcc
file wrongly lacking it?  Must it be given to copy code from a non-libgcc
file into libgcc?

-- 
Joseph S. Myers
jsm28@cam.ac.uk

^ permalink raw reply	[flat|nested] 9+ messages in thread

* Re: [3.2/3.3/HEAD] Make all the manuals unambiguously DFSG free
  2003-01-24  1:27 Richard Kenner
@ 2003-01-24  1:42 ` Joe Buck
  2003-01-24  1:58   ` Zack Weinberg
  2003-01-24  1:52 ` Joseph S. Myers
  1 sibling, 1 reply; 9+ messages in thread
From: Joe Buck @ 2003-01-24  1:42 UTC (permalink / raw)
  To: Richard Kenner; +Cc: zack, gcc-patches, gcc


> [ Zack's proposed change to GCC manual to make Debian folks happy ]
>     I am not interested in rehashing the debate, and since the motivation
>     for the change is to avoid controversy, assertions that there is
>     nothing wrong with the status quo will be taken as evidence in *favor*
>     of the patch, since they demonstrate the existence of controversy.
>     But if you believe that we *cannot* legitimately make this change then
>     please say so, giving justification.

Zack, try to watch your tone.  Language like "I am not interested in
rehashing the debate", or "assertions that there is nothing wrong with the
status quo will be taken as evidence in *favor*", come across as arrogant,
and this is not the first time that you have used such an approach.

Under the GCC Project's rules, demonstration of the existence of
controversy on a "political" matter isn't an argument in favor of a
change; just the opposite: it means that the SC must vote on the issue,
and it takes a 3/4 vote to make a change.  However in this case, even the
SC has no say, as Kenner points out:

On Thu, Jan 23, 2003 at 05:09:42PM -0500, Richard Kenner wrote:
> You cannot make that change because you are changing the terms of a copyright
> held by the FSF.  Only the FSF can legally make such a change.
> 
> This has nothing whatsoever to do with whether making such a change is
> or is not desirable.

Exactly.  The FSF will need to approve any change that changes license
terms or that, without their approval, would violate the license.

In any case, it appears that the FSF and the Debian folks may find
themselves at an impasse: the FSF wants to be able to spread its message
("propaganda" for those who don't like it) by including it in manuals, and
Debian seems to be trying to extend the DFSG to non-software and wants to
bar attempts to do this.

Nevertheless, if you can talk RMS into a resolution that both he and the
Debian folks can live with, more power to you.  Come back when you've
convinced him; I hope you're a patient man.

^ permalink raw reply	[flat|nested] 9+ messages in thread

* Re:  [3.2/3.3/HEAD] Make all the manuals unambiguously DFSG free
@ 2003-01-24  1:27 Richard Kenner
  2003-01-24  1:42 ` Joe Buck
  2003-01-24  1:52 ` Joseph S. Myers
  0 siblings, 2 replies; 9+ messages in thread
From: Richard Kenner @ 2003-01-24  1:27 UTC (permalink / raw)
  To: zack; +Cc: gcc-patches, gcc

    The executive summary is that there is consensus that the FDL meets
    the guidelines as long as there are no Invariant Sections, Front-Cover
    Texts, or Back-Cover Texts.  If any of these appear, the license is
    not agreed to be DFSG-free, and it has been seriously proposed that
    such documentation be removed from the Debian distribution.

    I believe we should proactively avoid this controversy.  Therefore,
    the following patch removes all Invariant Sections, Front-Cover Texts,
    and Back-Cover Texts from the manuals distributed as part of GCC.
    (The manuals untouched by this patch already have none.)

    I am not interested in rehashing the debate, and since the motivation
    for the change is to avoid controversy, assertions that there is
    nothing wrong with the status quo will be taken as evidence in *favor*
    of the patch, since they demonstrate the existence of controversy.
    But if you believe that we *cannot* legitimately make this change then
    please say so, giving justification.

You cannot make that change because you are changing the terms of a copyright
held by the FSF.  Only the FSF can legally make such a change.

This has nothing whatsoever to do with whether making such a change is
or is not desirable.

^ permalink raw reply	[flat|nested] 9+ messages in thread

end of thread, other threads:[~2003-01-24 11:15 UTC | newest]

Thread overview: 9+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
     [not found] <20030123150602.A15387@synopsys.comþ2>
2003-01-24 11:53 ` [3.2/3.3/HEAD] Make all the manuals unambiguously DFSG free Kai Henningsen
2003-01-24  2:01 Richard Kenner
2003-01-24  4:51 ` Joe Buck
  -- strict thread matches above, loose matches on Subject: below --
2003-01-24  1:27 Richard Kenner
2003-01-24  1:42 ` Joe Buck
2003-01-24  1:58   ` Zack Weinberg
2003-01-24  5:24     ` Joseph S. Myers
2003-01-24  1:52 ` Joseph S. Myers
2003-01-24  5:32   ` Joe Buck

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).