David Laight <david@l8s.co.uk> writes:
> However a 'cleanroom' implementation has to be done from
> documentation that is in the public domain (or some such
> definition).  Using GPL source code isn't...

The standard cleanroom techniques work just fine for GPLed source.
Group A reads the source and produces a functional specification, high
level description of the algorithms, etc., and group B implements the
replacement from the specification group A produces.  Nothing in the
GPL interferes with this process, nor is public domain required--it is
sufficient that the source has been published.

This is generally considered a very conservative procedure designed
to be resistant to legal challenges.  The same programmers reading the
original source and then reimplementing the algorithms can produce a
product that is not a derivative work in the copyright sense, but it
will generally be harder to establish this in court than it would be
with a cleanroom process.

Otherwise, BSD copyright *BSD wouldn't be possible, since many of the
routines written to remove all the USL intellectual property were
written by people who had read the original USL encumbered code (this
was the position of the AT&T/USL suit that tried to quash the release
of a USL-free BSD).

#include <stdianal.h>
-- 
"about 15 percent of the people are screwballs, lightweights and boobs
   and you would not want those people unrepresented in Congress."
	    -Alan Simpson, former US Senator (R, Wyoming)