Bug#218105: tetex-base: woeful copyright file

[frank@kuesterei.ch (Frank Küster)]

James Troup <james@nocrew.org> schrieb:

> Package: tetex-base
> Version: 2.0.2-5
>
> | Some individual parts of this distribution may have their own
> | copyright.  Please look into the respective files for their copyright.
>
> The copyright file should list the license of all files, not tell you
> to look at random other files.  [Yes, I know this is a pain for large
> packages and lots of people don't do it, but...]

I have thought about a feasible way to do this. I propose the following:

A. The format of the copyright file will be (after the general stuff and
   the license of the bundle+infrastructure itself) a list of license and
   the packages that use these, like this:

   LPPL:

   amscls, amsmath, amstex, babel,...

   GPL:

   ae, context, ...

   etc. With "package" in this context I mean what is in one directory
   on CTAN, e.g. the subdirectories of
   CTAN://tex-archive/macros/latex/contrib/. The files that belong to
   each of these packages would be listed in a separate file, to keep
   the copyright file readable. 

   Is this o.k.?

B. How to get the license information?

   I propose to rely on the CTAN Catalogue in the version packaged with
   tetex, insofar as it has sufficient information. i.e. if the
   Catalogue says "License: LPPL", the package will be listed under this
   license without further investigation. 

   I propose this because this saves a _lot_ of work. I tried to find,
   e.g., the license of the ams* packages, and there seems to be no
   information besides a very terse 4-line-copyright in each file, but
   it is listed as LPPL in the catalogue. I guess this is the result of
   a communication between AMS representatives and the CTAN masters.
   I trust the CTAN maintainers, and I don't want to repeat this
   communication (for this and probably a lot of other packages). 

   For common licenses (especially LPPL and GPL) I would only name them
   (and include one copy of LPPL), as indicated above. For other types
   of licenses, I propose to name the license in the copyright file and
   classify it (e.g. under "modification requires renaming" or the
   like), and then refer to the file (and location in the file) where
   the explicit license can be found. Otherwise, we'll have to gzip our
   copyright file, and it will be hard to browse through it.

   I propose to handle the CTAN "pd" license type (for Public Domain) in
   the same way as LPPL and GPL. I investigated a couple of these
   packages, and there was always only a short notice in the sense of
   "You may freely use, modify and distribute this code". This notice
   mostly was in each of the individual files, not in a central
   copyright/README file. I think it is not necessary and not useful to
   include links to each of these files in the copyright file, or even
   cite the respective sentences.

For packages where no common license, or no license at all, is indicated
in the packaged catalogue, I propose to resort to the following
resources:

- The current Online Catalogue, assuming that a license clarification
  also holds for the (possibly older) version shipped with tetex, or

- To the installed files, or

- As a last resort, communication with the authors.

By the way, I have asked Thomas Esser, tetex's upstream maintainer: He
doesn't have a list of licenses-by-package. He is content with knowing
that each of the files can be distributed (for free or sold), and he is
sure about that. He said there's some GFDLed stuff, so this will have to
be taken out (post-sarge), or we'll need a tetex-(doc-)nonfree again.

Any comments on that?

Regards, Frank
-- 
Frank Küster, Biozentrum der Univ. Basel
Abt. Biophysikalische Chemie