Re: ACM PDE Code GNU Compliant?

[John Weatherwax]

Hello,
   Sorry for the long lapse in time... I did contact the developer of 
PDECOL and he seemed interested in giving the code to octave (see below 
for the response (thanks Dr. Sincovec) but I am not sure what needs to 
be done legally.  Can anyone on this list point me in the right 
direction?  This may be covered in the FAQ's in which case I'll 
apologize right now and check there next but since the code is currently 
under the ACM copyright I wanted to check if anyone had any first hand 
knowledge of what specifically needs to be done to get code GPL'ed and 
included in octave?  For instance should I contact them, do they need to 
sign something, do the original authors need to sign something, if so 
what, etc.

   Thanks very much with you patience with a newby,

   John Weatherwax

John,

Sorry for the long delay in responding.  What do we need to do to 
release the code under the GPL?  I assume that this must also be done 
with the approval of ACM TOMS but have not pursued this.

I have included Niel Madsen on this reply.

Rich Sincovec

Richard Sincovec
Henson Professor and Chair
Computer Science and Engineering
University of Nebraska - Lincoln
256 Avery Hall
Lincoln, NE 68588-0115
(402) 472-2401


John Weatherwax wrote:

                Dr. Sincovec,

I am writing to tell you of an octave gateway that I've written
interfacing your very high quality code PDECOL.f.

http://www.mit.edu/~wax/Software/pdecol.html 
<javascript:ol('http://www.mit.edu/~wax/Software/pdecol.html');>

For solving one-dimensional PDES of the following form:

u_t = F(x,t,u,ux,uxx)

This will enable a user to define a few simple octave functions (like
the right hand side of F above, the initial conditions, the boundary
conditions etc and the gateway routine uses the C++ features of octave
to pass the FORTRAN pointers to the FORTRAN functions needed to
perform its integration.  This is analogous to how Matlab handle ODEs
in its functional arguments.  There is an example of the octave call
reproducing the first example from the paper on the web sight.

This gateway and the combination of a Matlab like environment, enables
very quick modeling turn around: from problem formulation to solution
visualization in seconds.  I see this as a powerful tool for students
and researchers.  I would like to make this tool available to as wide
an audience as possible.  I wanted to write to inform you of its
presence and to make a request.

The current license for the software PDECOL.f is with the ACM.  As
such, the existing copyright states that the use of the software is
restricted to non-commercial/academic purposes only, which is TOO
stringent for full inclusion into octave.  Octave requires a more
liberal license governed under the General Public License (GPL).  As
such, it cannot be included in octave.  If you (as one of the original
authors) were to release the code under the GPL this code could be
included in the core octave sources potentially reaching thousands of
users and having a much greater impact.  I am writing to see if you
would consider doing this.  It would be a potentially very influential
way to get quality mathematical software into much greater use.

I hope that this email is not to preachy.  I have found the pdecol
routines code interesting and enjoyable to work with and would truly
like to see more high quality software for PDEs available to non
specialists.

Thanks very much for your time,

John Weatherwax

PS. As other additional projects, I am planning on writing gateways
for other high quality software in TOM's (some of which you have
written).

PPS. I was unable to locate Dr. Madsen on web searches, if you could
forward this along to him I would be most appreciated.



David Bateman wrote:

> John W. Eaton a écrit :
>
> > On 13-May-2005, John Weatherwax wrote:
> >
> > |     1)  Would this be helpful?
> >
> > Yes, it would likely be useful to have PDE software for Octave.
> >
> > |     2) Could I use the ACM FORTRAN code (pdecol.f) in octave, i.e. 
> > is it | GNU compliant?  I am attaching a link to the ACM software 
> > copyright and | license information...I'm not very good at reading 
> > these things. | |     
> > http://www.acm.org/pubs/copyright_policy/softwareCRnotice.html
> >
> > This license is not compatible with the GPL because it restricts the
> > use of the software to non-commercial/academic purposes only.
> >
> > |     I think the BLAS algorithms originally appeared in ACM TOMS so 
> > I am | hoping that this will be alright.
> >
> > I don't think the BLAS was ever covered by this license.
> >
> > jwe
> >  
> The presence of an ACM license is not necessarily a problem as the 
> ownership of the code remains with the author, and they can release 
> the code to you under a different license.. So if you like the code, 
> try contacting the author and see if they will release it to you under 
> the GPL.
>
> Cheers
> David

--

John L. Weatherwax

Lincoln Laboratory, Massachusetts Institute of Technology
244 Wood Street
Lexington, MA 02420-9108 USA

Phone : 781-981-5370
Fax   : 781-981-4739
Email : address@hidden

%--