Re: problems of the fdl

>>>>> "Eli" == Eli Zaretskii <[email protected]> writes:

    Eli> If you do deprecate the FDL and stop using it, what is the
    Eli> alternative?

We don't use the FDL yet in XEmacs core.  We already have a couple
of packages using the FDL for their docs, but they're almost all
integrated into GNU Emacs.

This license stuff is an enthusiasm-killer.

    Eli> The GPL doesn't seem appropriate, as documentation is not
    Eli> code, and so ``free software'' is not a good description of a
    Eli> free manual.  The language of the GPL doesn't fit such uses
    Eli> of it.

The Creative Commons ShareAlike license that Alex favors looks good to
me, except that the warrantee that you've vetted all content for
copyright permissions seems awfully risky for a wiki.  There are
several variants, I think this is the one Alex likes:

http://creativecommons.org/licenses/by-sa/1.0/

However, for Emacsen, especially XEmacs, pymacs, etc, it suffers from
the killer flaw that it isn't the GPL.

    Eli> Do we even have a good alternative?

The GPL is ugly but legally clean for this purpose; you just have to
add some definitions (of terms like "source and object code") to the
license statement.  The only real problem with the GPL is that it
doesn't inhibit DRM.  But all alternatives to the GPL are worse
because they are incompatible with the GPL.  Dual licensing has its
attractions, but it's not possible to enforce dual licensing on
downstream distributors (Catch-22, the clause that says you must
dual-license makes it incompatible with the GPL because the GPL, of
course, doesn't have that clause).

The long-run solution, to my mind, is for the FSF to revise the GPL to
look a lot more like the CC SA.  In particular, it has a much better
anti-DRM clause than the one in the FDL.  While somewhat awkward to
refer to code, the idea that "running a program" is "a digital
performance" of the program resonates with me; I like it.  It unifies
the code and documentation ideas AFAICT.  And it would have the killer
advantage of being a version of the GPL, and therefore already
compatible with the huge body of copylefted software that allows
distribution under current or later versions of the GPL.

The CC SA as such has a potential flaw in that I it may mean "external
deployment" (eg, simply offering an application service on the web)
triggers the source distribution.  Many people like this, but rms has
declared it non-free in the form used by Apple Public Source License.
The Open Source License http://www.opensource.org/licenses/osl.php
gives an explicit definition of External Deployment.  I don't know
what rms (and/or the FSF/Eben Moglen) have to say about the OSL.

    Eli> There's also a more general issue, already mentioned in this
    Eli> discussion: any license restricts freedom in some ways.

Sure.  But we've all come to terms with that.  Some people like
restricting the "bad guys'" options, some people want to pay due
respect to the wishes of the giants whose shoulders give such a great
view.

The problem is that using multiple strong copyleft licenses restricts
freedoms that everybody agrees are important.  Do you realize, Eli,
that without the explicit permission of the FSF you personally are not
free to combine the Emacs sources and the Texinfo manual into a
Knuthian "literate programming" document, and distribute it?  It's not
even obvious to me that the FSF (assuming that rms has as much say in
this as I think he does) would grant permission to do so.  I think rms
probably would, but I'm not sure.  (What license would he suggest?)

Wouldn't that be a wonderful, wonderful thing: a document which
simultaneously can be compiled into Emacs, processed into an Info or
printed User's manual, or processed into a textbook on all the
algorithms and techniques needed to implement a powerful customizable
editor?  There is something seriously wrong when that vision is made
technically illegal.

That also answers your comment that code isn't documentation.  Of
course it is.  It always has been, otherwise we'd still be writing in
machine language.  And documentation is code.  What else is the
INSTALL file, but a program to drive the SysAdmin wetware?  Of course
I'm stretching the point, but the "literate programming" problem is
real, I think.  It's not just TeX, anymore.  The annotated Linux
sources have been published as a book.  And it's just a matter of time
before somebody decides to do it with Emacs, IMO, and probably GCC.
Apache and GNOME and ....  I wish I were a CS undergrad! :-)

-- 
Institute of Policy and Planning Sciences     http://turnbull.sk.tsukuba.ac.jp
University of Tsukuba                    Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
               Ask not how you can "do" free software business;
              ask what your business can "do for" free software.