Brad Salai wrote:
> Colin Wood wrote:
> 
> <snip>
> 
> > So, you basically cannot use Linux kernel source
> > code in a NetBSD kernel or vice-versa.  For the mac68k ports of Linux and
> > NetBSD, this means that each camp must figure out all the hardware for
> > themselves, duplicating a lot of effort in some cases.  Of course, some
> > sufficiently kind kernel hacker could fully document the hardware outside of
> > the source code, but that almost never happens ;-)
> 
> This  is my first thought on this, and I might change my mind after
> thinking about it some more, but I disagree in part.

feel free, i'm certainly not a lawyer :-)
 
> Say for example that some Linux code, possibly a header file had
> information about a bunch of registers that you needed to make some
> hardware work. You can take the information, the location of the
> registers, and anything else factual, from the Linux code and use them
> to write NetBSD code without infringing the copyright on the Linux code,
> as long as you don't take anything expressive, which would include
> comments in most cases, and any actual code that was present (although
> you can take the ideas included in the code). 
> 
> What is sometimes done, is one person looks at the copyrighted code, and
> writes a spec that includes all the factual stuff, but none of the
> copyrightable expression (the actual code), and another person who
> hasn't seen the original code uses the spec to write functionally
> equivalent code, which does not infringe, even if it turns out to be
> very similar, or even identical. This is because identity isn't enough
> for there to be copyright infringement, there has to be copying. The
> presence of a lot of identical code is evidence of copying, because
> coincidence only goes so far, but it is only a rebuttable presumption.
> If you can prove independant creation, there is no copyright
> infringement.

oh certainly.  this is more or less the "clean room" approach which has
been used for years.

> The leading case on this involved copying a telephone directory. The US
> Supreme Court decided that the information (names, addresses and phone
> numbers) in a telephone directory could be copied without infringing,
> even though it took a lot of work to assemble it. They said that these
> were pure factual things, not subject to copyright. They threw out the
> so called sweat of the brow test which previously had given protection
> to things that took a lot of work to assemble. 

that's hilarious.  i'm surprised no encyclopedias were ever involved in
cases like this :-)

> Since Colin does this for a living

well, not really.  i validate processors for a living.  i just play around
with netbsd on the side when i have time ;-)

> I expect he knows all of this, (the
> reference to the mythical kind kernal hacker (KKH) writing the spec is
> what I was talking about)  but I wanted to clarify a little. The KKH can
> use the Linux code to document the hardware, it doesn't have to be the
> person who wrote the code who does it.

of course!  however, i think that there is a rather fine line between
using someone else's code as documentation and wholesale copying from it.
i'd hate to be the person who has to prove that my code was original work
given that it looks just like this other guy's code i've looked at ;-)

later.

-- 
Colin Wood                                 cwood@ichips.intel.com
Component Design Engineer - PMD                 Intel Corporation
-----------------------------------------------------------------
I speak only on my own behalf, not for my employer.