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

Interpreting the GFDL GR

After going over the debate leading up to the vote, I think that the
most honest and least polemic interpretation of the winning GR would
be as follows:

 a) The effect of the GR is to change our standards of interpreting
    license terms for a piece of software that include a clause that
    appear to aim at preventing a licensee from circumventing a
    requirement to distribute source by distributing source in a form
    that it would be particularly cumbersome or illegal for a
    recipient to use according to the freedoms otherwise provided by
    the license.

 b) When determining whether software is DFSG-free, the clauses
    described by (a) shall not be interpreted according to their plain
    English meaning. Instead we will form an conservative guess about
    what the licensor is likely to have _intended_ by the clause. In
    forming this guess we will employ extra-textual knowledge about
    the licensor's policies and the circumstances of choosing or
    drafting the license. We will then determine whether the
    _intention_ of the licensor was to deny a licensee any rights
    required by the DFSG.

In particular:

 c) The so-called "DRM" clause of the GFDL which literally reads
         You may not use technical measures to obstruct or control
         the reading or further copying of the copies you make or

    shall, for the purpose of the DFSG, be interpreted to mean

         You may not distribute copies in a form or under such
         circumstances that the intended recipient cannot exercise the
         rights given to him by this License without violating the
         United States DMCA or similar laws existing elsewhere.

 d) When the GFDL requires a licensee to

         include a machine-readable Transparent copy along with
         each Opaque copy

    we shall, for the purpose of the DFSG, assume that this
    requirement is implicitly qualified by

         If distribution of Opaque copies is made by offering access
         to copy from a designated place, then offering equivalent
         access to copy the Transparent copy from the same place
         counts as distribution of the Transparent copy, even though
         third parties are not compelled to copy the Transparent copy
         with the Opaque one.

    (I am not sure whether ftpmasters should apply this interpretation
    when deciding whether they dare put binary packages containing
    opaque copies of GFDL-licensed works on our ftp servers given the
    current pool infrastructure. I would suggest not. However, that is
    not directly a DFSG issue).

I assert that this interpretation is most faithful to the arguments
presented by proponents of Amendment A during the discussion. In
particular, when confronted by arguments that a literal reading of the
GFDL leads to nonfreedom, proponents of the winning text did not
generally challenge the nonfreedom of the literal reading, but instead
argued that we should not use the literal reading because that was
"obviously" not what the FSF meant when drafting the license.

In particular as regards the DRM clause, I don't think it is logically
possible to claim that the *literal* meaning of the sentence allows
transmission in encrypted emails or storing on a multiuser machine
with mode 004. And nobody appeared to make that claim; instead they
claim that the sentence should not be read literally in the first

And once we stop reading literally, we might as well be open about
jumping ship and instead trying to guess at the licensor's intentions.
Thus I don't think we should contort ourselves to try to shoehorn as
much as possible of the GFDL *wording* into a DFSG-free
interpretation, such as the quasi-literal interpretations suggested by
Anthony DeRobertis.

Further, as far as I can make out very few people disagree that the
interpretations I give in (c) and (d) above are probably fairly
accurate descriptions of the FSF's intent - even we who maintain
that this is not what the license _says_ are willing to believe that
it is what the FSF _tried to say_ but failed. Similarly everybody
seem to agree that if the GFDL actually did say what I write in (c)
and (d), those two points would not be freedom issues.

We could go all the way and start interpreting _all_ of all licenses
according to the licensor's intent rather than the literal meaning of
the license text, but I think that would be dangerous and extremely
undesirable, so we should not start doing that _in general_ withtout
_explicitly_ being instructed by the Developers By General Resolution.
Therefore my proposal is to narrow the licensor's-intent principle to
clauses of the general kind that are problematic in the GFDL. The
description in point (a) above is my best attempt to define such a
"general kind" of restrictions without being too much a GFDL-specific

Henning Makholm                                   "Det er trolddom og terror
                                                         og jeg får en værre
                                               ballade når jeg kommer hjem!"

Reply to: