Is documentation different from software [Re: Proposed statement wrt GNU FDL]
On Sat, 26 Apr 2003, Matthew Palmer wrote:
> I was about to pipe up with "but we don't distribute novels with Debian"
> until I realised that we want to distribute a few other novel-like things -
> pure documentation not associated with a specific software program (eg the
> hoary old chestnut of the RFCs).
And interactive fiction, and screensaver artworks, and actual novels
> So, I have to start thinking again about
> documentation in a different light to software.
I'm still completely unconvinced that there is any useful distinction
between a fractal generator, the generated fractal image (with human input
on parameters, colors, etc), a paper describing what fractals are, a
document describing how to make pretty fractals from the generator, and a
poem inspired by a fractal.
They're all encodable as streams of bits. They all require effort and
creativity to create. They all are useful to a recipient. They all are
free iff a user can modify and copy them.
> Could we produce a distinction amongst our offerings in the following
Why do we want to produce a distinction where there is none?
> DFSG applies to software. (duh). DFSG applied to docs for software (for
> all the very good reasons given in the draft Debian GFDL FAQ). But for
> "other documentation" (and we'd need to have a pretty clear definition of
> what that was before going too far forward) there could be a second set of
> guidelines for it's inclusion in Debian.
Why? Why would we call a poem which cannot be modified "free" if we don't
also think a poetry generator which cannot be modified is free?
> To roast a hoary chestnut, I've not yet seen a good argument why we'd want
> the RFCs to be relicensed as DFSG-free, apart from the "so it can go into
> Debian main".
Oh, that's easy. I wish they were free so I could reuse the good bits in
other works, and otherwise improve my world by creating and distributing
This is absolutly no different from why you'd want a piece of software to
be licensed freely.
> Modifying an RFC and re-releasing it is not a good thing
It's exactly as good a thing as modifying Apache and re-releasing it.
The modified version is expected to fit someone's needs better than the
> The DFSG, social contract, GNU Manifesto, software
> licences, all of these are non-software documentation which we want to
> distribute (OK, maybe not the GM) but which in general we don't *want* to be
The owners of these documents should be encouraged to make them free, and
if not, they should go in non-free.
License texts themselves are a special case. The actual terms of the
grant of permission cannot be changed, and can only be encoded in the
original wording. That makes perfect sense, and is fundamentally
different from a copyrightable work in a definable way.
Mark Rafn email@example.com <http://www.dagon.net/>