Re: Inconsistencies in our approach
Lynn Winebarger <email@example.com>
>It can also be turned around - why claim everything is software except
>to force DSFG restrictions where they are unnecessary or undeserved?
One good definition of software is "the part of a computer that's not
hardware". Another is "Information in a format designed
to be read by a machine". It's hardly artificial to use these
definitions and say that everything Debian distributes, except the
physical CDs, is software.
Anyway, nobody's trying to force DFSG restrictions where they are
The point has already been made that the DFSG requirements *are*
just as necessary for documentation as they are for
programs. (The same motivations apply.)
The point has further been made that the requirements are just
as necessary for standards documents as they are for programs. (The
same motivations apply.)
I have to add a new section to my FDL FAQ. Here's a draft version:
It's not about misrepresentation!
A lot of people seem to think that the FDL's "invariant
sections" are there to prevent people from misrepresenting other
people's opinions. This is totally incorrect.
Suppose you (Mr. Foo) write an essay: "Why the BSD license is best, by
Mr. Foo." No matter what copyright license your essay is under -- even
if it's in the public domain -- nobody can modify it to "Why the GPL
license is best, by Mr. Foo." That's fraud (misrepresenting your
opinion). It's quite likely libel in many jurisdictions as well. It's
certainly illegal and unethical. It has absolutely nothing to do with
Similarly, if someone modified a copy of "The GNU Manifesto" to say
something else, it would be fraud to distribute the modified copy under
the title "The GNU Manifesto". (It's a distinctive title with a
distinctive meaning.) Actually, GNU used in this fashion is probably a
trademark as well, so it's probably a trademark violation too.
If you want to add a clause to a copyright license which says:
"This license shall not be construed to allow anyone to misrepresent my
opinion, or to commit fraud, or to deliberately mislead anyone,"
that's perfectly free. This is conceptually similar to clauses
requiring modified programs to prominently note that they have been
modified, which many free licenses require. This is *not* the nature
of GFDL "Invariant Sections".
If you want to make *absolutely sure* nobody mistakes the modified
version for your original essay, you can put a clause in the license
All modified versions must prominently state that they do not
necessarily represent the opinions of Mr. Foo.
This might be a little irritating, but it's certainly free. GFDL
"Invariant Sections" go much further than this.
If your essay is in the public domain, or under a license which allows
free modification, someone else can use your words in the essay "Why the
GPL license is best, by Mr. Bar." Note that you are *not* being
misrepresented. GFDL "invariant sections" prohibit this.
Perhaps you care about being identified as the author of the original
essay. In this case, you probably want to require Mr. Bar to write in
his modified version something like:
This uses some material from 'Why the BSD licence is best', by Mr.
Foo, but does not necessarily represent Mr. Foo's opinions.
It's perfectly reasonable to require this in all modified copies, and
plenty of free licenses have such requirements. The GPL requires
crediting the original author. This is not what GFDL "invariant
sections" are for, either.
GFDL'ed invariant sections prohibit modification. You probably just
want to require that modified versions don't pretend to be your version,
and possibly that they credit you as the author of the original. This
can be accomplished much more easily, with better, freer licenses.
Nathanael Nerode <neroden at gcc.gnu.org>