Hannu Krosing <hannu@tm.ee> writes:
> I know that Compaq was forced to do a clean-room re-engineering of PC
> BIOS
> (two teams - the dirti one with access to real bios athat does
> description
> and testin and the clean team to write the actual code so that they can
> prove they did not "steal" even if the result is byte-by-byte simila)
> for
> similar reasons
Compaq wasn't forced to do this. They did it on the basis that
following this complex procedure would guarantee that they would win
if it came to a court case (which it did not). But there is no way to
tell whether a simpler procedure would not win in court.
The GPL is a copyright license. Copyrights, unlike patents, only put
limitations on derived works. If you independently write the same
novel, and you can prove in court that you never saw the original
novel, you are not guily of violating copyright. That's why Compaq
followed the procedure they did (and it's why Pierre Menard was not
guilty of copyright infringement).
But that's novels. As far as I know, there is no U.S. law, and there
are no U.S. court decisions, determining when one program is a
derivative of another. If you read a novel, and write a novel with
similar themes or characters, you've infringed. However, there is
less protection for functional specifications, in which there may be
only one way to do something. When does a computer program infringe?
Nobody knows.
Ian