Re: Can we freely use AMOS in Octave?

[Steven G. Johnson]

Robert T. Short wrote:]
> The "Amos" code was obtained from netlib, not the ACM.  Here is what 
> netlib says about restrictions on use:

It doesn't matter where we obtained it.  What matters is who holds the 
copyright to the code (if any) and under what terms they allow 
distribution (if any).

> This algorithm is a package of subroutines for computing Bessel
>  functions and Airy functions.  The routines are updated
>  versions of those routines found in TOMS algorithm 644.
>
> So this really isn't the TOMS code anyway. 

This is not how copyright law works.  If ACM holds the copyright to the 
TOMS code and forbids distribution under the GPL, then that restriction 
also applies to any derived work of the TOMS code.  Netlib clearly 
states that AMOS is derived from the TOMS code.


> Furthermore, this code was 
> published elsewhere (see the references in Amos' code). 

This is not how copyright law works.   Just because they published the 
code elsewhere in addition to ACM doesn't mean that they didn't transfer 
the rights to ACM.

> I really don't 
> think the ACM owns the copyright to the code we are using, and I think 
> the only person that can grant or revoke permission is Amos himself or 
> more likely Sandia Labs. 

When you publish in ACM TOMS, normally the author assigns the copyright 
to ACM.  See: http://toms.acm.org/AlgPolicy.html

	"Authors of copyrightable algorithms (or their employers) are required 
to transfer the copyright to ACM upon acceptance of the algorithm for 
publication, in accordance with ACM policy to own copyright on ACM 
published material."

The ray of hope here is that Sandia also published the ACM code in 
SLATEC, where they explicitly stated that it is in the public domain. 
In my understanding, this means one of three possibilities:

a) Sandia had released the code into the public domain, i.e. 
relinquished copyright, prior to its publication in ACM TOMS.  Hence, 
there was no copyright to assign to ACM.

b) Sandia made some other arrangement with ACM that allowed them to 
release AMOS into the public domain in SLATEC.

c) Sandia was mistaken: they had forgotten about the copyright transfer 
to ACM when they claimed to release SLATEC into the public domain.

Although (c) is a theoretically possibility, we have no reason to 
believe that Sandia did not perform due diligence when they released 
SLATEC.  It seems reasonable to take the word of the original authors 
that the work is in the public domain (unless and until ACM sues Sandia 
over SLATEC).

Steven