Re: The debate on Invariant sections (long)
Richard Stallman <rms@gnu.org> writes:
> Your message repeated over and over that you think the GFDL isn't
> free, but didn't even try to justify that claim. I continue to
> believe that the GNU FDL is a free documentation license.
>
> The key question is: is the FSF prepared to abandon its use of
> non-free licenses for manuals?
>
> That question is like "Will you stop beating your wife?" All it
> proves is that you are willing to sink low. I'm not going to discuss
> the issue with you.
I think I didn't understand your position before. I was assuming that
everyone thought that the invariant sections were non-free. I'm
surprised that you don't think so, but I now understand it, where I
didn't before.
I was confused because in the free-doc.html article you say "I don't
believe that it is essential for people to have permission to modify
all sorts of articles and books". That makes it sound like you think
the right to modification isn't important: you do think it could
exist, and that it's a freedom, just that it isn't important.
Now you have clarified, apparently, that you think it really isn't a
freedom at all.
The FSF's definition of freedom includes the ability to modify.
That's a crucial freedom, and it's precisely that freedom which is
impinged in the case of invariant sections.
Now you might say it's an unimportant freedom, in which case I am
happy to explain the cases where it does raise a real barrier.
I had thought you were saying "that's an unimportant freedom". Now it
appears that I misunderstood, and you are actually saying "that's not
a freedom at all".
The key freedom is fourth freedom in the list of freedoms at
http://www.gnu.org/philosophy/free-sw.html. It's the freedom to
"improve the program".
Now the problem with Invariant Sections is that they prohibit the
modification of certain political statements.
Right away we see the problem: the FSF has of course occasionally
altered its own political statements as it sees how to word things
better, or takes account of changed realities. So it is clear that
such statements *do* have room for improvement. This means that it is
logically possible to improve such sections: they are not already
maximally perfect.
So this means that an Invariant Section is something which *could* be
improved, but thanks to the license, it cannot be. This is exactly
the sort of thing that Freedom Four is designed to address.
Consider a software author who says "I want to prohibit modifications
because I only want you to use this program to speak the standard
protocol. You can trust me to authorize updates when the standard
protocol changes." We would regard this as non-free. The author
knows that updates might improve the program, but has decided that the
risk of bad updates outweighs the risk of good updates, and so he
prohibits them all.
Similarly, the FSF's Invariant Sections prohibit all updates because
of a fear that certain kinds of bad updates will happen. But Freedom
Four doesn't care about this sort of risk: the whole point is that
users are not subject to a copyright owner's judgments about which
updates are good and bad--*even* in the case that the user and the
owner completely agree.
There are other kinds of improvements that can be made. One example
is taking a free manual and turning it into dynamic online
documentation (much like Lisp docstrings). The rules now force the
entire invariant section to be distributed anytime even a single
docstring is distributed which was copied from that manual.
As the above mentioned URL says: "Being free to do these things means
that you do not have to ask for permission." Since it is clear that
Invariant Sections can be improved, it is clear that they fall within
the ambit of the Fourth Freedom: "The freedom to improve the program."
Since the freedoms must be held without the need to ask or pay for
permission, a document which uses Invariant Sections cannot be free.
Note that the Fourth Freedom does not distinguish between different
kinds of improvements; it does not care whether the improvement is
"important"; it does not care about the risk of harmful changes.
Now the article on free documentation says that you don't think
permission to modify political articles is important. So be it--but
the point is that the requirement to *distribute* them along with the
non-political documentation is a problem; it functions as a
restriction on the modification of the *technical documentation*.
Your apology in the free-doc.html article was added only after Debian
people started complaining about the FSF's practice; but what you say
there still holds in part: "It must be possible to modify all the
*technical* content of the manual." The presence of Invariant
Sections in fact makes it impossible to do certain kinds of
modification of the technical content: importantly, it makes
impossible those which take the content and turn it into something
completely non-manual-like and non-book-like entirely.
Thomas
Reply to: