Re: Patch for Copyright Notices in maintain.texi [WAS: Re: Approving Dezyne, savannah task 16067]

[Alfred M. Szmidt]

   >        notice in each nontrivial file in the package.  As a rule of thumb,
   >        any interesting file more than ten lines long should be considered 
and
   >        could be nontrivial for this purpose.  This includes header files

   It seems to me, if a file isn't interesting, it just doesn't belong
   in the package; if a rule (of thumb) is referred to, it may make sense
   to explain it.

Maybe the wording here, "interesting", is what is the issue.  I think
what Janneke was trying to convey (correct me if I am wrong), was a
"interesting" file in the sense that it is copyrightable, and not
e.g., autogenerated, just some boiler plate, simple test cases, etc.

   >        If a header can be added without any complications that is always
   >        preferred; please make a balanced choice here.  The important thing 
to
   >        keep in mind is that there cannot be any discussion about the
   >        copyright and license status of any file.  For example, a whole
   >        directory tree with a similar layout such as a test suite with
   >        generated baseline data should probably use one toplevel 
@file{README}
   >        file describing how to generate the test baseline data, instead of
   >        hundreds of similar or identical @file{README} files.

   In such case, if a small part of the testsuite is copied, there can be
   a discussion about the copyright and license status of any file; I think
   the concept of the balance adds uncertainty because the relative values
   of clear legal status vs. developer's convenience will be subjective.

Not sure I follow, could you elaborate?

   >        Another example is small code snippets that are included verbatim in
   >        the (texinfo) documentation and can also be processed directly by 
the
   >        program being documented.  Because the example is included as an
   >        integral part of the documentation, it is covered automatically by 
the
   >        copyright and license of the manual.  Adding a @file{README} file in
   >        this case is encouraged.

   I believe this isn't consistent with the GPL HowTo
   <https://www.gnu.org/licenses/gpl-howto.html#why-license-notices>:

           If a release has one statement that “This program is
           released under license FOO,” in a central place such as the
           README file, that makes the situation clear for that
           release. However, programmers often copy source files from
           one free program into another. If a source file contains no
           statement about what its license is, then moving it into
           another context eliminates all trace of that point. This
           invites confusion and error.

This is about copying from one program to another, unrelated program.

The case Janneke raises above is within the same program, where
snippets of code are verbatim included into a manual which already has
a license chapter or similar.  So, such snippets if copied from the
source tree would have the license of the program or manual depending
on things -- since that has a central place where the license is
mentioned.