[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Gnu-arch-users] [OT] Debian should move all GPL sw to non-free

From: Andrew Suffield
Subject: Re: [Gnu-arch-users] [OT] Debian should move all GPL sw to non-free
Date: Sun, 28 Mar 2004 15:01:53 +0100
User-agent: Mutt/

[WARNING: This mail does not constitute guidelines for writing
DFSG-free licenses, and compliance with the points advanced here does
not guarantee the resulting license will be DFSG-free]

I'm afraid I'm going to have to play the 2 of trumps.

"G" in DFSG stands for "guidelines". Trying to apply it like a program
is the mistake of OSI. There's some flexibility here.

Also <cough> we've done this one on debian-legal before. Actually the
"GPL is non-free" argument is a perennial source of amusement, and
it's been a long time since we saw a new one.

On Sat, Mar 27, 2004 at 08:38:22AM -0800, Tom Lord wrote:
>   The Debian position on the DSFG is that it permits no restriction
>   whatsoever on modifications to the substantive content of a free
>   work.  The sole exception is, of course, those kinds of minimal
>   restrictions which are necessary to implement the copyright
>   ownership and licensing which protects the freeness of a work.

This isn't really accurate, although it's close. Shuffle the words a
little and you've got it:

"it permits no restriction on (any modifications whatsoever)"

More on this later.

>   By extension, any license which imposes content restrictions, other
>   than those which implement minimal requirements of copyright law,
>   must not be DSFG-free.
>   The GPL is such a license.
>   In section 2c, the GPL says:
>     c) If the modified program normally reads commands interactively
>     when run, you must cause it, when started running for such
>     interactive use in the most ordinary way, to print or display an
>     announcement including an appropriate copyright notice and a
>     notice that there is no warranty (or else, saying that you
>     provide a warranty) and that users may redistribute the program
>     under these conditions, and telling the user how to view a copy
>     of this License.  (Exception: if the Program itself is
>     interactive but does not normally print such an announcement,
>     your work based on the Program is not required to print an
>     announcement.)
>   I call your attention to the list of things which must be printed or
>   displayed, in particular:
>       that users may redistribute the program under these conditions
>   Such a statement in the start-up message of an interactive program
>   is not legally necessary.   It is political speech.

Here's where I start to substantively disagree. It *may* be political
speech, but it does not *have* to be.

>   It is political speech that the FSF typically includes in its
>   interactive programs.   For example:
>         % gdb
>       GDB is free software and you are welcome to distribute copies
>         of it under certain conditions; type "show copying" to see the
>         conditions.  There is absolutely no warranty for GDB; type
>         "show warranty" for details.
>         GDB x.y, Copyright XXXX Free Software Foundation, Inc.
>         (gdb)
>   Note the phrase "you are welcome to distribute copies of it under
>   certain conditions".  There is no legal necessity for that phrase
>   outside of the requirements of GPL.   
>   You are free to modify GDB.  You are free to change the wording of
>   that start-up message.  However, you are not permitted to change the
>   wording of the start-up message in a way that causes it to fail to
>   explicitly inform users that they may redistribute GDB and directing
>   them to the GPL to see the conditions of that redistribution.

This is distinct from the GFDL, where you are not free to change the
wording. The dividing line for DFSG-freedom lies between these two.

Section 2c of the GPL is a perfect example of how to write an
acceptable restriction on modification, and it is really close to the
line (the closest such restriction that exists and is still DFSG-free,
to my knowledge). The restrictions imposed by the GFDL are
unacceptable because they fall short. The GFDL neither permits removal
(even via hoop-jumping) nor rephrasing, even to correct mistakes.

We routinely accept clauses of the form:

"You must visibly acknowledge the following groups and individuals"

We routinely reject clauses of the form:

"You must include the following acknowledgement in the following
"You must visibly include this HTML code in any web pages generated by
 this program"

And we tell the license authors to rewrite the clause to be like the
first one.

The "History" section also fails the Chinese Dissident test;
distribution of a modified version must not require disclosing your

We have tried to open a dialog with the FSF to get the offending
clauses of the GFDL adjusted. RMS refused to discuss the issue; some
others did not, and that's still progressing (painfully
slowly). Invariant sections will never be acceptable, but there's no
reason (other than RMS) why the GFDL can't be "Free so long as there
are no invariant sections or cover texts".

>   The same brute-force application of the DFSG that leads to the
>   conclusion that GFDL is non-free must also reach that conclusion
>   that GPL is non-free.

Now it becomes clear why you took this line of reasoning. Here's where
you went wrong while working it out:

This was not a blind or brute-force application of the DFSG. The DFSG
is merely a convinient way to talk about freeness issues; it is not a
function for testing for freeness (the real function for this is "ask

We started out from "Clearly it is possible to apply the GFDL in a way
that results in non-free works. Under what circumstances does it
result in free works?". The first cut at this classified the FSF's
application of the GFDL as being DFSG-free.

Then we looked more carefully at these cases. After a while, a range
of real, practical issues were uncovered, of the form "The GFDL as
applied here prohibits these actions", and prohibiting these actions
was determined to be non-free. Therefore, this application of the GFDL
was determined to be non-free.

At this point we believed that there were still applications of the
GFDL which were free - those which had no "Invariant sections", no
"History" section, etcetera.

Finally we noticed[0] some more mundane issues with the GFDL; these are
the sort of problem that many new licenses face. These render all
applications of the GFDL non-free. To the best of our knowledge they
are simply bugs, but RMS refuses to talk about anything relating to
the GFDL. At one point (after repeated attempts to get something done
about these bugs) he acknowledged one, claiming it had never been
mentioned before, while ignoring the other which was presented in the
same mail, but said that attending to it would have to wait until
after the GPL v3 was prepared. covers the
details, including the "other" issues.

Ironically, most of these issues were raised either by us or by
others, including some of the bugs, during the "consultation" period
for the GFDL, and all of them were ignored. We're not sure who was
being consulted or why.

[0] Actually one of these had been raised before, but it got lost
    while we were chasing invariant sections.

  .''`.  ** Debian GNU/Linux ** | Andrew Suffield
 : :' : |
 `. `'                          |
   `-             -><-          |

Attachment: signature.asc
Description: Digital signature

reply via email to

[Prev in Thread] Current Thread [Next in Thread]