wfmath_1.0.1-1_amd64.changes REJECTED

[Stephen M. Webb]

On 08/05/2013 11:47 AM, Paul Tagliamonte wrote:
> On Mon, Aug 05, 2013 at 11:41:34AM -0400, Stephen M. Webb wrote:
>>> // The original code included the following notice:
>>> //
>>> //     When you use this, send an email to: matumoto at math.keio.ac.jp
>>> //     with an appropriate reference to your work.
>>
>> The source file in question, wfmath/MersenneTwister.h, is authored by Richard J. Wagner <rjwagner at writeme.com> and is
>> distributed under the BSD 3-clause license.  It is an extension of the original implementation of the Mersenne Twister
>> algorithm by Shawn Cokus <cokus at math.washington.edu>.
> 
> (therefore a derived work)

Yes, indeed.  The Wagner work (the one in question here) is derived from the Cokus work.  The Cokus work contained the
following historical note in commnets:

  // The code as Shawn received it included the following notice:
  //
  //   Copyright (C) 1997 Makoto Matsumoto and Takuji Nishimura.  When
  //   you use this, send an e-mail to <matumoto at math.keio.ac.jp> with
  //   an appropriate reference to your work.

Note that the Cokus work is an original work implementing the algorithm published by Matsumoto and Nishimura but
containing none of their code (which can be found at [2].  Notice the disclaimer at the top of [3].  It is possible, if
you squint and hold your head just right, that this wording could be interpreted as a license restriction on the Cokus code.

When Wagner published his version he included the historic note in a comment near the top of his source file, in a
separate stanza from the license statement.  He changed the wording of the note:  the antecedent of the pronoun can only
refer to the Matsumoto and Nishimura code at [2] because the Cokus code contained no such clause, only the historic note
documented above.  In addition, Wagner credits copyright to the algorithm authors:

  // Copyright (C) 1997 - 2002, Makoto Matsumoto and Takuji Nishimura,

This is interesting because the 2002 version of Matsumoto and Nishimura [4] contains only this clause.

   Any feedback is very welcome.
   http://www.math.sci.hiroshima-u.ac.jp/~m-mat/MT/emt.html

It is clear that when he published his code, Wagner was aware that the original email notice was no longer a license
condition of the Matsumoto and Nishimura code.  There is good reason to believe he intended it as a historical note
rather than a license condition, then antecedent of the pronoun in the clause referring to the obsolete 1997 version of
the Matsumoto and Nishimura code rather than the clearly unrestricted 2002 version referenced in the copyright.

> Again, this does not read as a request.

Fair enough.  It is possible to interpret the Wagner clause as a de novo non-free distribution license condition rather
than an upstream requirement.  Given this particular verbatim source file has been used in dozens of source packages in
the Debian archive (for example, php5 and blender) [5], if this particular package is rejected on the grounds that its
source is non-free, should I start filing bugs against all the other packagaes in the archive that it turns out are
non-free so they can either be fixed or removed?

[1] The original has disappeared, but a copy can be found at http://dirk.eddelbuettel.com/code/octave-mt/cokus.c.txt
[2] http://www.math.sci.hiroshima-u.ac.jp/~m-mat/MT/VERSIONS/C-LANG/980409/mt19937int.c
[3] http://www.math.sci.hiroshima-u.ac.jp/~m-mat/MT/VERSIONS/C-LANG/ver980409.html
[4] http://www.math.sci.hiroshima-u.ac.jp/~m-mat/MT/MT2002/CODES/mt19937ar.c
[5] http://codesearch.debian.net/search?q=MersenneTwister.h

-- 
Stephen M. Webb  <stephen.webb at bregmasoft.ca>

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 899 bytes
Desc: OpenPGP digital signature
URL: <http://lists.alioth.debian.org/pipermail/pkg-games-devel/attachments/20130809/f48d2714/attachment.sig>