Free documentation licenses

[Tom Hull]

John Cowan wrote:
> 
> On Wed, 29 Nov 2000, Nelson Rush wrote:
> 
> > The GPL shouldn't be used for documentation, it is intended for use with
> > software. I think RMS one time agreed with me on this when it came up in one
> > of my other projects. Which is why he created the FDL, which is specifically
> > designed for documentation.
> 
> Be that as it may, if you publish a book which contains full source with
> detailed commentary (like the TeX and METAFONT sourcebooks or the
> Lions Book), the book is going to be a derivative work of the
> program.

Scott Maxwell's "Linux Core Kernel Commentary" seems to argue otherwise.
This book (published by Coriolis) contains a very large extract of the
Linux source code (license GPL), followed by a short commentary (copyright,
all rights reserved). I don't know what Coriolis's thinking was, but I can
imagine two plausible justifications:

 1) The book itself is an aggregation. (The book also provides a CDROM, so
    the GPL source code is also available in usable form, which is after all
    the primary concern of GPL.)

 2) The GPL license has the side effect of significantly expanding the bounds
    of fair use -- primarily by making it hard to argue that any amount of
    fair use quotation economically impacts the copyright holder.

I would think that an in-line commentary could not be justified as an
aggregation -- that this would depend on fair use. I also think that
extended fair use is consistent with the intent of GPL, which is to
keep the source code public and freely accessible and reusable.

TeX and Metafont are slightly different cases, since the doc is woven into
the program source -- effectively, the doc and source are one. Nonetheless,
the published books (my copies are Addison-Wesley, 1986) have conventional
all rights reserved copyrights. The sources, of course, are not GPL, and
even if they were the copyright holder would be entitled to do this.

> Do you suppose that if you got ahold of the Windows
> source code and published it in such a book that Microsoft wouldn't
> scream "Copyright violation!", and rightly so?

Depends on the license. If there is no license, all rights are reserved,
so one would have to depend on fair use. ESR has claimed that his
publication of Microsoft's "Halloween" documents is defensible under
fair use.

> What is, or is not, a derivative work is a matter for the law
> (ultimately, the courts), not the wording of any license, and the
> GPL goes out of its way to say so.
> 
> Very simply:
> 
>         book containing full source is a derivative work of the source
>         source is under the GPL
>         GPL says derivative works must be under the GPL too
>         --------------------------------------------------------------
>         book can only be published under the GPL

Maxwell's book, noted above, is a counter-example.

> None of this applies to a book that merely comments on the program,
> or uses at most fair-use excerpts from it, like the Linux kernel book.

Maxwell's quotation is 39338 lines of code.

-- 
/*
 *  Tom Hull * thull at kscable.com * http://www.ocston.org/~thull/
 */