Begin forwarded message:
From: "Fitzell, Julian" <jfitzell@cincom.com> Date: October 12, 2010 10:00:56 AM GMT+02:00 To: <bonzini@gnu.org>, <stephane.ducasse@inria.fr> Cc: <BSchwab@anest.ufl.edu>, <paolo.bonzini@usi.ch> Subject: Re: [Pharo-project] License question
Are you sure? I've heard opinions expressed that looking at a library and copying the API might constitute a derivative work.
I'm not certain if that's correct or not but if you're purposely trying to skirt the terms of the GPL, you might want to make certain.
Julian
----- Original Message ----- From: Paolo Bonzini <paolo.bonzini@gmail.com> To: Stéphane Ducasse <stephane.ducasse@inria.fr> Cc: Fitzell, Julian; Wilhelm K Schwab <BSchwab@anest.ufl.edu>; Bonzini Paolo <paolo.bonzini@usi.ch> Sent: Tue Oct 12 08:47:56 2010 Subject: Re: [Pharo-project] License question
On 10/12/2010 09:35 AM, Paolo Bonzini wrote:
On 10/11/2010 11:29 PM, Stéphane Ducasse wrote:
Let's go forward in time. NativeBoost, Alien or FFI provide callbacks, I do some adaptation on my end, and out comes a pretty capable wrapper for the GNU Scientific Library. That sentence trivializes a LOT of work :( Details aside, is the wrapper code itself affected by GPL
Yes.
That said, there is a way out. You could have a simple-minded library which has the same interface as the bindings but is written in pure Smalltalk. Then, the same thing can run fast under the GPL, or slow under MIT license.
I meant to do this exactly for GSL for a long time, making it compatible with the code from "Object-Oriented Implementation of Numerical Methods" so that it could act as a fallback. However, I never even started it.
http://www.amazon.com/Object-Oriented-Implementation-Numerical-Methods-Intro...
Paolo