License of Debian-specific parts in packages, generally and in particular

[Frank Küster <frank@debian.org>]

Hi,

in particular, tetex-base has a woeful copyright file (#218105), and
while I'm trying to resolve this, I came across the fact that some of
the Debian-specific code (maintainer scripts, templates,...) does
not have a license statement. The maintainer scripts don't even have a
proper copyright statement, i.e. one has to guess from debian/changelog
who has contributed what.

More generally, I found out that this is the case for many packages
(just a random pick: emacs21{-common}, kdebase-bin, scigraphica) have
the same deficiencies. An example for a "good" package is the xfree
Packages; furthermore, in xfree86-common/copyright, the copyright for
all Debian-specific contributions is assigned to SPI.

Therefore, I would like to raise some questions. If they have been
discussed before, please give me a pointer.

1. Shouldn't we add a note to the Policy (or the Developer's Reference)
   that there should be a license statement for the Debian-specific
   parts in debian/copyright? I think we should, and it should be a
   "must" directive post-sarge.

2. Should we encourage maintainers and contributors to assign the
   copyright to SPI, as the x people did?

3. Is there any advice on whether to put the debian-specific part under
   the same license as the upstream work, or whether this does not
   matter?

4. How should we proceed with old contributions? Especially if
   maintainers have frequently changed, or complex patches from the BTS
   have been applied, it might be hard to find out all the copyright
   holders. 

   I think one can assume that anybody contributing code [1] to a Debian
   package is willing to put it under a DFSG-free license, but one
   cannot guess at all which license this should be. Therefore a
   transition strategy could be made that would allow old code with
   unknown or unreachable authors in the package if it is marked a such,
   but require a rewrite if substantial changes have to be made anyway.

Regards, Frank

[1] at least if the code is complex enough to warrant a copyright at all.
-- 
Frank Küster, Biozentrum der Univ. Basel
Abt. Biophysikalische Chemie