[GRASS-dev] r.terraflow license

[Glynn Clements]

Brad Douglas wrote:

> > I suspect that there's a reason why, of the many clauses in the GPL,
> > the FSF's recommendation is to include the disclaimer of warranty in
> > each file.
> 
> I had asked about this some time ago when I was updating library
> "header" text.  Here's the text that was settled on:
> 
> "This program is free software under the GNU General Public License
> (>=v2). Read the file COPYING that comes with GRASS for details."
> 
> Since the file 'COPYING' comes with GRASS, the user can refer to that
> for details of the license.

Right; but the disclaimer of warranty isn't an issue of copyright. We
don't need to draw the user's attention to the licence; copyright law
puts the burden of obtaining consent on anyone modifying or
distributing the work. Claiming not to have noticed a particular
restriction wouldn't free the distributor from it.

OTOH, if there is a need for the developer to disclaim liability, the
burden would be on the developer to inform the user of this. Failure
to adequately inform the user could result in liability which would
otherwise have been excluded.

If you check the source files for any GNU software, you will normally
find that the user is referred to the [L]GPL for copyright issues, but
the disclaimer appears directly in the source file, even though it can
also be found in the licence to which the user has just been referred.

> However, we may want to augment the
> statement to say "(v2)" instead of, "(>=v2)".  GPLv3 exists and there
> has been no consensus on using it (although, the GRASS statement implies
> we already do).

If we were to remove the "or any later version" text, we would quickly
forfeit the right to subsequently adopt a later version, as any
contributions would be assumed to be licensed under "v2 only" terms. 
If we ever wanted (or needed[1]) to adopt a later version, we would
have to obtain explicit consent to do so.

[1] It's probably only a matter of time before e.g. readline starts
being distributed under v3. If we're bound to v2-only terms, providing
binaries could quickly become troublesome.

-- 
Glynn Clements <glynn at gclements.plus.com>