Re: FDLed manuals

[Nathanael Nerode <neroden at twcny dot rr dot com>]

>So to me the GFDL is "restrictive" in the same way that the GPL is
>restrictive because you're not allowed to remove the GPL notice
>from source code.

Interesting point of view.  It's wrong, of course, since a Manifesto is 
not even similar to a copyright notice. :-)  I have zero objections to 
the GPL-like requirements of change notices and copyright notices, or to 
the Endorsements rules, all of which are entirely reasonable.

Obviously people can put their manuals under whatever licence they like.
This is just a major, major tactical error on the FSF's part, promoting 
a non-free licence as 'free'.

Here's my favorite of the (many) problems with Invariant Sections.
Consider the definition of "Secondary Sections".  Either it
means only "any section designated secondary", or it creates endless 
problems for derivative works.  Suppose a manual is about 
topic A and contains an "Invariant Section" about topic B.

Suppose you then make a derivative work which is about topic B.  You 
cannot delete the Invariant section, according to the original 
manual's licence; applying the FDL to the derivative manual, it cannot 
*be* an Invariant Section, so the derivative cannot be licensed.  So the 
presence of an invariant section on topic B actually prevents the 
creation of derivative works on topic B entirely!

There is no length limit on Invariant Sections.  Provided that the 'main 
subject' has sufficiently many pages to be the the largest section, 
Invariant Sections could easily occupy, say, 3/4 of a manual.

In the case of derivative works, it's quite easy to end up saddled with
useless Invariant Sections.

For a more realistic example, suppose I make a program which was, in 
some ways, compatible with GNU Emacs.  I want to use the appropriate
sections of the GNU Emacs manual to document the work-alike parts.
Of course, I have to carry along the GPL (even if I'm not licensing my
program under it), the GNU Manifesto, and the "Distribution" section
describing the distribution rules for GNU Emacs.  Well, maybe it's 
better if I write my documentation from scratch.  And if I do 
that, I'd better not look at the Emacs manual at all so that I won't be 
in danger of being sued by the FSF.

Still consider the FDL free?  If so, you have a mighty odd definition of 
free, but I won't argue further.  Seems more like proprietary to me.

>As always, Debian is free to go elsewhere if they don't like GCC.
>There's always LCC!
>
>Stan

No, Debian likes GCC.  What's at issue is the manuals, and only the 
manuals.

What will actually happen, in all probability, is that Debian will ship 
GCC but not the GCC manual, and similarly for all the other GNU products.  So
much for the intent of promoting free, reusable documentation. :-(

I wouldn't be in the least bit surprised if a project was started to 
write alternate manuals for GCC & company under an unambiguously 
DFSG-free license.  Which would be a horrible duplication of effort,
but hey, the current GCC manual isn't that good anyway, so maybe it 
would be for the better.

--Nathanael