Re: GPL and linking (was: Urgently need GPL compatible libsnmp5-dev replacement :-()
We actually seem to be getting somewhere, slowly.
On 5/7/05, Raul Miller <moth.debian@gmail.com> wrote:
> On 5/7/05, Michael K. Edwards <m.k.edwards@gmail.com> wrote:
> > On 5/6/05, Raul Miller <moth.debian@gmail.com> wrote:
> > > I think this "attempts to" quip is meaningless.
> >
> > How would you like me to say it? "Purports to"? "Professes to"?
> > "Makes an honest but flawed effort to"?
>
> That would depend on the meaning you are trying to convey, wouldn't it?
All of the above. The license text, as actually written, contains an
erroneous statement:
... a "work based on the Program" means either the Program or any
derivative work under copyright law: ...
(So far so good; that is the formula, in legal documents as in
mathematics and other formal styles of discourse, for declaring a
variable and binding it to a set. Ends in a colon, which is a fairly
strong separator in formal English.)
... that is to say, ...
(Paraphrase alert! Signals the start of a comment, extending to the
end of the line! Grammatically, an adverbial phrase, with its scope
indicated by parallelism in the sentence structure, which in this case
is at the level of noun phrases.)
... a work containing the Program or a portion of it, either verbatim
or with modifications and/or translated into another language.
(A noun phrase, which the drafter of the GPL may well have thought was
equivalent to the noun phrase to which it is equated, but which is in
fact not. I repeat not. Did I say not? Thought I did. Google the
flippin' Berne convention and read Article 2, inclusive of headings.
Show me a contradiction in recent case law at the appellate level from
a developed country and I'll unbend a little -- but not much.)
As the GPL is an offer of contract, not a statute, the common law
principle is to construe it against the offeror, and to accept the
definition given above -- if the judge considers it credible that the
licensee intended to accept that definition. Which, for the record,
is the only definition I have accepted since I started wearing long
pants.
It is indeed the case that, in construing a _statute_, a common law
court is obliged to find if possible a construction which gives every
word meaning. If this were a statute, and if it were not otherwise
absurd to do so, the court would perhaps be obliged to find that "work
based on the Program" was meant to include the larger scope given in
the paraphrase, especially if it could be determined from the
legislative record that the consensus in the legislative body at the
time was that that was what they meant. For better or for worse,
neither RMS nor Eben Moglen nor the FSF nor Debian is the legislative
body for a competent jurisdiction. :-)
Want the first two sentences of GPL Section 0 in (more or less) set
theory terms?
U \equiv \{ x : "x is any program or other work which contains a
notice placed by the copyright holder saying it may be distributed
under the terms of this General Public License" \}
DEFINITION: choose P (the "Program") \element U
E \equiv \( \{ P \} \union \{ y : "y is any derivative work of P under
copyright law" \} \)
DEFINITION: W ("works based on the Program") \equiv E
C \equiv \{ z : "z is a work containing the Program or a portion of
it, either verbatim or with modifications and/or translated into
another language" \}
Hypothesis: W = C; FALSE (C \superset W)
By (common) law: W remains defined as E.
(Note that \equiv is TeX for the three-bar-equal-sign "define" symbol
and that \superset is meant to be the one with the little "not equals"
underneath it.)
And as for the more pejorative implication of "purports" and
"professes", it may or may not have been an honest error in
understanding at the time the GPL was drafted, but I simply don't
believe that the FSF doesn't know better by now. I think it may well
have been an honest mistake in the original, but remember, it was
originally drafted to cover Emacs and the GNU alternatives to
command-line Unix programs, and I don't even think Unix had dynamic
linking back then, so they probably didn't think too hard about it,
assuming that the minimal scope of a "work" extended to the bounds of
an executable anyway (which isn't even necessarily true).
I feel for them, really I do, but somewhere along the way Eben Moglen
got a law degree and a law chair at Columbia and I'm sorry but this
stuff does kind of come up in law school. Somewhere along the way,
the FSF FAQ has gone from being wrong to being, let's say,
disingenuous. If they want the GPL to say what the FSF FAQ says it
does, then they can "fix" it in GPL v3, relicense the upstream of the
projects on which they hold copyright and anyone else they can
persuade, and we'll see which fork survives.
> If there's nothing of relevance to say (which is what I believe is the case
> here), then the best way to convey that lack of meaning is through not
> saying it. Alternatively, if there is a flaw, ou could succinctly describe
> the flaw. Or, if succinctness is impossible you could give a url for a
> page which describes the flaw.
>
> I've yet to see anything which makes me believe that the GPL
> carries this flaw you're suggesting.
Relevance is in the eye of the beholder. As is succinctness. And
making someone believe something is not really within my powers.
> > Do you not understand my interpretation that the use of quotes
> > around "work based on the Program" means that the writer is
> > defining it as shorthand for "either the Program or any derivative
> > work under copyright law"?
>
> No, I did not understand that that's what you thought the quotes
> meant.
>
> Now that you've explained that, I understand that that's what you
> think the quotes mean, but I disagree with you.
>
> The quotes mean that the phrase has a special significance.
> You've given a valid definition of that phrase, but that's the
> definition of the whole phrase, not the quotes.
I don't understand this last sentence. Would you mind stating your
reading in more or less formal terms?
> > And that an attempt is then made to paraphrase (restate, whatever)
> > the latter phrase, and that restatement is just plain wrong?
>
> You've yet to show why the restatement is wrong.
How about now?
Cheers,
- Michael
Reply to: