commons-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Sébastien Brisard <>
Subject Re: [math] Porting someone else's work (license and all that)
Date Thu, 29 Sep 2011 04:43:19 GMT
Good news! Please read below Pr. Saunders's answer. Should I send the
ICLA through the usual way (secretary@ if I remember well), or shhould
I use another channel, explaining precisely to the recipient at Apache
what it is all about?

Dear Sébastien,

Thank you for your interest in SYMMLQ and for explaining the background
in detail.  Certainly I'm glad for you to make use of the code in this way.
I'm sorry I haven't done anything to the Fortran and Matlab versions for a
long time.  Our SOL website points to python versions in case they are of
interest.  MINRES and MINRES-QLP would also be suitable for indefinite
systems (and sometimes preferable -- e.g., if you wish to terminate early).
(We have a new f90 implementation of MINRES-QLP.)

I'd be glad to hear a little about your materials application in case a
particular feature suggests a preferable solver.

The CPL was rather arbitrary.  It's fine to use the Apache license too.
A signed copy is attached.

Best wishes for this endeavor,

Michael Saunders
Professor (Research)
Stanford University

Le 29 septembre 2011 06:38, Sébastien Brisard
<> a écrit :
> Hi Phil,
> no Java code available. I'm polishing up my port, will submit it in
> the coming days (weeks?) for you to review
> Sébastien
> 2011/9/28 Phil Steitz <>:
>> On 9/27/11 11:58 PM, Henri Yandell wrote:
>>> CPL is a weak-copyleft license - ie) putting source derived from it
>>> inside Commons Math will make all of Commons Math CPL licensed.
>>> If Mr Saunders is the only copyright owner, and were happy to sign a
>>> grant or ICLA, then the CPL wouldn't apply in our use of it (ie: no
>>> issue).
>> Right, probably a software grant would be appropriate.  Is the Java
>> code available anywhere that we can look at, or are you talking
>> about creating a Java port?
>> Phil
>>> Hen
>>> 2011/9/27 Sébastien Brisard <>:
>>>> Dear Commons-Math community,
>>>> I've recently been using quite extensively the SYMMLQ linear iterative
>>>> solver proposed by Page and Saunders, for its ability to handle
>>>> symmetric *non-definite* systems. I know it's not the only one around,
>>>> but that's not the point.
>>>> My point is I would like to commit this implementation to the CM
>>>> library. However, the core of the code is a mere port of the FORTRAN
>>>> code downoadable on Pr. Saunders's homepage
>>>> The FORTRAN
>>>> source comes with a license, which I include here for the sake of
>>>> completeness
>>>> {quote}
>>>> symmlq_f77.README
>>>> The software for SYMMLQ (f77 version) is provided by SOL, Stanford University
>>>> under the terms of the OSI Common Public License (CPL):
>>>> 11 Feb 2000: First set of files available for download from SOL.
>>>> Please send comments to Michael Saunders, SOL, Stanford University
>>>>                650-723-1875
>>>> -----------------------------------------------------------------------------
>>>> The f77 version of SYMMLQ involves the following files:
>>>>   symmlq.f
>>>>   symmlqblas.f   (not needed if you have BLAS-1)
>>>>   symmlqtest.f
>>>> {quote}
>>>> My question is: how do I proceed from here? Is the CPL compatible with
>>>> the Apache License (I think it is)? Courtesy would probably require to
>>>> get in touch with Pr. Saunders. Do you allow me to do it on behalf of
>>>> the CM community? Maybe one of the historical committers would rather
>>>> do it? If Pr. Saunders agrees on signing the Commons License
>>>> Agreement, would it still be required that the CPL be included?
>>>> Thanks for your answers,
>>>> Sébastien
>>>> ---------------------------------------------------------------------
>>>> To unsubscribe, e-mail:
>>>> For additional commands, e-mail:
>>> ---------------------------------------------------------------------
>>> To unsubscribe, e-mail:
>>> For additional commands, e-mail:
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail:
>> For additional commands, e-mail:

To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message