[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Amendment to GR on GFDL, and the changes to the Social Contract



On Sat, Feb 11, 2006 at 10:52:01AM +0000, Roger Leigh wrote:
> Henning Glawe <glaweh@debian.org> writes:
> 
> > On Fri, Feb 10, 2006 at 07:58:52PM -0500, Glenn Maynard wrote:
> >> This really just isn't a problem that needs fixing.  Once in a while, you get
> >> confused or desperate people on d-legal trying to argue "we allow license
> >> texts to be unmodifiable, so this invariant ode to my cat should be allowed,
> >> too!", but you can't stop those stupid arguments by changing the DFSG.  You
> >> just end up replacing one dumb argument with another, equally dumb argument,
> >> and complicate the guidelines in the process.
> >
> > just one thought: we have programs in main, where derived works are
> > only allowed as original source+patches (TeX comes to my mind...)
> > couldn't it be basically the same thing with GFDL documents? if
> > there is an invariant section with an 'ode to my cat', why can't we
> > add a section to the document telling the 'ode to my cat' is bloody
> > stupid. this would be in some sense equivalent to a patch, only the
> > interpreter is not the computer but the human brain (which is the
> > target architecture for documentation anyways).
> 
> It's not equivalent.  A patch /changes/ the original to give you
> something new, whereas adding additional material merely /extends/;

a 'patch' in the first run is also an extension to the original source;
only an interpreter (in most cases, /usr/bin/patch) makes a 'change' from it.

in the case of documentation, where the 'interpreter' is the combination of
eyes/ears and parts of the brain, a sentence telling 'the following section
is bogus' can be seen as a patch.

my point is: when it comes to documentation, how do you classify source and
binary. One could argue the postscript/pdf/html file on the disk (or even
it's dead-tree edition) is only an intermediate product, like e.g. java
bytecode; the binary form is the knowledge a person absorbs when reading the
document. It's all a matter of interpretation.

> it's not hard to see long-term maintenance problems with this.  See
> the debian-vote archives for more detail.

ask people working on software with 'derived-works-only-as-patches' licenses
;)

-- 
c u
henning

Attachment: signature.asc
Description: Digital signature


Reply to: