Re: Results for Debian's Position on the GFDL
On Tue, Mar 14, 2006 at 09:29:40PM -0500, Raul Miller wrote:
> On 3/14/06, Glenn Maynard <glenn@zewt.org> wrote:
> > (I don't think any special attempt to prevent the technical measures
> > themselves are necessary, since the GPL's source requirements already
> > did that: an encrypted, locked, unmodifiable copy is not source.)
>
> Ok, but the legal right to modify a work does not mean that you have
> the practical ability. More to the point, the GFDL prohibits the use
> of technical measures to enforce any of the more obnoxious clauses
> of the GFDL.
I can't think of any way that one could use "technical measures" to
fulfill the GPL's requirements without providing usable source. That
is, you can't DRM the the source, since--I'd imagine--nobody can actually
edit the source in that form. (Or, if you could, then the encryption
keys needed to do so would effectively become part of the source.) If
"technical measures" are applied to the source to restrict it, it's
probably not the source anymore, by the GPL's definition.
(Speaking here only of the DRM subtopic. You could lack the practical
abliity to use the source if it's in a weird language, but that's the
"transparent copy" topic--let's keep these things separate for sanity ...)
> All you need is a broadly available shim -- for example, something to convert
> word format to xml and xml back to word, to make it straightforward to
> modify the word document in a generic editor.
So you can use any format, all you have to do is reverse engineer it,
find an open format that's a complete superset of it (if one exists),
and write a two-way lossless converter? That seems tantamount to
a prohibition--especially if, like many writers, you're not even a
programmer. (I am, and that's still a daunting task.)
> No one (you included) has even cared to identify subsets of packages
> where this is true. And we have some fairly broad subsets where code
> is re-usable.
I don't claim that everything must be compatible with everything else.
(Though I think it's a good goal, which is why I stick to the MIT license,
to maximize compatibility.) With ordinary licenses, at the very minimum,
a work's license is compatible with the license itself (any others are a
bonus). Patch clauses don't do that. They're compatible with nothing.
> I don't believe I've ever heard of anyone even suggesting that someone's
> right to modify software should be revoked because their changelog
> entries (or whatever) are legally ambiguous.
The GFDL specifically says that it must "clearly and legibly identify you".
Ambiguity and clarity are opposites, and pseudonyms do not identify you.
--
Glenn Maynard
Reply to: