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

licensing of cephes library (affects labplot, grace, ?...)


Please CC replies to me, because I'm not on the list.

I noticed a problem with the license of the cephes library, which is included in the Debian packages labplot (maintained by me) and grace (maintained by Torsten Werner), and possibly others. At least python2.2-scipy and python2.3-scipy seem to have some cephes stuff in them, though I haven't looked at their sources. Both labplot and grace are GPL.

In the version of labplot and grace that are in Debian, cephes appears to be unlicensed. Looking for versions of cephes on the web, I've found variations on the following statement:

  Some software in this archive may be from the book Methods and Programs
  for Mathematical Functions (Prentice-Hall, 1989) or from the Cephes
  Mathematical Library, a commercial product. In either event, it is
  copyrighted by the author.  What you see here may be used freely
  but it comes with no support or guarantee.

This is all the licensing that seems to be around for cephes, and it does not seem to be DFSG-free as it stands, because it doesn't specifically allow modification and redistribution.

Torsten Werner and I have contacted the upstream authors of both labplot and grace, and it seems that they do not have any clearer idea than we do of the licensing for cephes. So I have emailed the author of cephes to ask him for clarification of the cephes license and whether he would consider relicensing cephes in a way that is both GPL compatible and DFSG-free. I specifically suggested the MIT or GPL licenses as a possibility, with the MIT license being possibly preferable, because it is simpler and easier to understand. He seems to be helpful - he responded quickly and seems to be wanting to help. But he doesn't appear to want to relicense cephes in that fashion for his distribution of it on his website [1]

The most recent communication from him follows below [2]. I am confused about whether this arrangement would be OK from Debian's point of view. Specifically, I would like to know whether,

1) if the labplot and grace upstreams put the proposed boilerplate into the cephes files in their distributions of labplot and grace, could Debian package and distribute labplot and grace with those cephes libraries included?

2) if the labplot or grace upstream were taking a long time to do this, would it be OK to put the boilerplate into the Debian distributions of those cephes libraries in the meantime, as a quick fix to make labplot and grace actually distributable?

I don't particularly want to hassle the cephes author further, since he is trying to help but he doesn't seem to want to change the way he mostly distributes his things. So if this solution is OK, I would like to go with that. But I want to be sure that it is OK for Debian first.



1.  http://www.moshier.net/

This is the most recent email from the cephes author, in repsonse to my suggestion that he consider releasing cephes from his website under the MIT license.

-------- Original Message --------
Subject: Re: licensing of cephes
Date: Tue, 28 Dec 2004 12:36:15 -0500 (EST)
From: Steve Moshier <steve@moshier.net>
To: Helen Faulkner <helen_ml_faulkner@yahoo.co.uk>
CC: twerner@debian.org, Ben Burton <bab@debian.org>, evgeny@stambulchik.net, Stefan Gerlach <gerlach@mbi-berlin.de> References: <41D14354.4040506@yahoo.co.uk> <Pine.LNX.4.58.0412280836250.9987@moshier.net> <41D17BE3.3020708@yahoo.co.uk>

I think you do not have to dictate the terms under which I distribute
the material through other channels.  You probably need only a
permission that allows you to distribute it under your particular
terms.  Here is an example boilerplate that worked for BSD.  A recent
go-around with the FSF resulted in something a bit more wordy but
similar in spirit.

   * Cephes Math Library Release 2.8:  June, 2000
   * Copyright 1984, 1995, 2000 by Stephen L. Moshier
   * This software is derived from the Cephes Math Library and is
   * incorporated herein by permission of the author.
   * [standard BSD license here]

Reply to: