Re: digital images

On Sun, 30 Jan 2000, Karsten M. Self <kmself[_at_]ix.netcom.com> wrote:
>
> In either event, NeXT would have been compelled to contribute their
> code or remove the GNU gcc code. LGPL allows works which utilize
> the library only in linking to be considered independent.
> Modification of the library itself produces a derivative work and
> is subject to the relicensing requirement:

The thing they would have done is make a library out of the main gcc code, then written a front-end that linked to the main library. Then (under the LGPL) they could keep their front-end proprietary.

> Claiming authorship to programmatic outputs of one's own program, when
> run by someone else on inputs of their creation, seems to fall outside
> the scope of copyright: "original works *of authorship* fixed..."
> (emphasis added). The authorship requirement appears to be missing
> in this case on the part of the compiler writer as applied to compiler
> outputs. Rather the compiler is generating a mechanical translation of
> a work, presumably of authorship, written by the programmer of the now
> compiled work.

What's troubling me is "mechanical translation". There is considerably more at work in a compiler than there is in, say, "translating" an ASCII code into a character on a screen. In particular, the compiler is generally an encoding of the author's knowledge of how to take the high level language and put it into a machine level form that is reasonably efficient. You could hire the compiler author to do it by hand and probably get marginally better or even worse code (and a lot more errors in the execution). There's a lot of creative content in the resulting assembly program that is purely due to the creativity of the compiler writer. I'd say the assembly code is a joint work of the programmer and the compiler writers.

Lynn

Lynn Winebarger
<owinebar[_at_]free-expression.org> Received on Mon Jan 31 2000 - 22:44:31 GMT