[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: What kind of copyright/license statement is required here?

Michael Tautschnig <mt@debian.org> writes:

> PS.: Please CC me, I'm not subscribed.


> I'm the maintainer of the sat4j package, which, in the upstream SVN,
> includes lots of test data (plain text files, think of them as
> mathematical equations). Currently these lack any kind of
> copyright/license statement, so I remove them before building the
> tar ball. It would, however, be desirable to package them as well,
> and upstream is willing to do whatever is needed have them included.
> So, my question is:
> - What kind of copyright/license statement is required? Should the
> usual (in this case Eclipse Public) license header be added, even
> though this is not "source code"?

The general principle (in my understanding) is: Make explicit and
unambiguous, to anyone who may receive the work from any source, the
copyright status and grant of license for every file comprising the
work, where:

* "copyright status" means which specific persons hold copyright, and
  in what years (since copyright is held for a limited time, at least
  in principle).

* "grant of license" is an explicit statement from the copyright
  holder of the form "you may do these things, normally reserved to
  the copyright holder, with this specific work, if you abide by these
  terms and conditions".

The grant of license may list all the terms, or it may refer the
reader to a separate document where the license terms are contained;
but without the "you may" and the "this specific work" statement, mere
inclusion of a license document is *not* a grant of license.

> - Should be it be added to each and every file?

The how-to-use-this-license-for-your-own-work part of the GNU GPL,
which speaks about program source code, recommends that both of these
should appear as a header block in every file. This makes it easier to
move such files from one work to another without losing their
copyright and license information.

That's good advice for program source code. However, it's only one way
of meeting the general principle of "make the copyright and license
explicit and unambiguous for every file".

Another possible way is to put all the information in one document
that specifies every file comprising the work, and gives the complete
set of copyright holders and which files contain their particular
work, and what years. This seems more attractive to some, since all
the information is in one place, and has an appearance of less
"clutter" in the work.

However, it makes the very dubious assumption that this information,
in its separate file, will continue to be accurately maintained as the
work inevitably accumulates new contributors, new copyright years, or
component works under different licenses. If that maintenance is not
done, the provenance of each work becomes that much harder to
determine. It also depends on that separate file accompanying any
individual file if it is redistributed separately from the rest.

For works in a form that don't have a clear place to put such a
"copyright header" (e.g. an audio recording, or a formatted data
file), the "copyright information in a separate file" approach makes
more sense; but the maintenance and aggregation burden is not lessened
in any degree.

 \         “Pinky, are you pondering what I'm pondering?” “I think so, |
  `\         Brain, but what kind of rides do they have in Fabioland?” |
_o__)                                           —_Pinky and The Brain_ |
Ben Finney

Reply to: