Re: SCO laying an ambush for the GPL?

[Chris Croughton]

On Mon, 19 Jul 2004 10:22:04 +0100, Nick Kew 
   <nick@hugin.webthing.com> wrote:

> Ref: SCO Reply Memorandum Re Discovery - published at Groklaw
> 
> The issue now seems to be "IBM has improperly contributed its own code to
> Linux in violation of a contract". Since it is not in dispute that IBM has
> contributed code to Linux, that would seem to boil down to an argument over
> whether SCO has rights over any or all of that code on the grounds of being
> a derivative work.

This isn't news, that's what it has been for some time.

> Now if I find my GPL code distributed with the copyright removed,
> I know there's a violation. And that applies also to derivative works
> under the GPL. So if I *suspect* violation I may need discovery to
> determine whether it is a derivative work under the GPL.

Would that give you a right to go poking around on the suspect's
machines?  I could say that I suspect you of using my code without
attribution (it's BSD-type licence, but that still requires that you
credit my code), but that shouldn't give me any right to go poking
through your code just "on suspicion".  It seems to me that 'discovery'
in this sense is another word for 'fishing' (as in "I don't know whether
the suspect has illegal drugs, but I want a warrant to go and search for
them just in case").

> The direction this is going seems likely to set a precedent that will be
> cited in any major GPL trial under US law. So if SCO isn't granted all
> the discovery it's asked for, that'll be a useful precedent for, say,
> M$ to deny discovery when someone accuses them of GPL violation.

In the SCO case, they are claiming that their code is in an open-source
product.  They don't need to 'discover' anything, they can use grep on
the source code of Linux and determine for themselves where their code
is used.  If it isn't published, whether IBM 'donated' it is irrelevant
because it wasn't used (and they won't be able to 'discover' such a
donation anyway).  Their attempt at 'discovery' it just a ploy to draw
the case out longer, they have already been required to say exactly what
they claim has been "improperly contributed" and have failed to do so.
If anything has been "improperly contributed", they should be able to
point to it in the published source code of Linux (back versions are
available from way before they claim the code was "contributed").

There's also the point that it's very hard to prove that a piece of code
has been copied, in most cases (some is of course so specialised that it
is unlikely to have been independently concieved, and some companies
don't even bother to file off the serial numbers).  Take the original,
remove all comments, rename all identifiers (to your own coding
standards, for instance), reformat the code, reverse some conditions,
move cases in switch statements around, and it will be difficult to
convince anyone that it's the same.  Of course, unless the code is
really specialised that may well be more work than writing it from
scratch...

(note followup)

Chris C