[linux-elitists] SCO Teleconference transcript (Friday, May 30, 2003)

Karsten M. Self kmself@ix.netcom.com
Sun Jun 1 21:46:36 PDT 2003


on Sun, Jun 01, 2003 at 08:58:04PM -0700, Nick Moffitt (nick@zork.net) wrote:
> begin  Karsten M. Self  quotation:
> >   - SCO claims that the answer to the question: is there evidence of
> >     infringing SCO code in the Linux kernel is "a resounding
> >     'Yes!'", with "hundreds of lines" of allegedly infringed code to
> >     be shown.  Much is made of the scale of the source code search:
> >     "we're dealing with multiple versions of Linux over many years,
> >     and huge amounts of source code".
> 
> 	Does this seem just the slightest bit innumerate to anyone
> else here?  I've been smirking at that "hundreds of lines" comment
> since I read it on this list a while back.  

Yes.

On the other hand, legal cases have swung on far less.

The Sega v. Accolade case (plaintif lost) was based on 42 bytes (plus or
minus a few) of data at the start of a ROM card necessary to get a game
to play (the ruling was based on the principle that copyright protection
may not be functional).  Other cases have involved little more than a
few lines of code in a larger codebase.

Your general point is valid, though it has to convince the court (jury
and/or judge, depending).  I suspect IBM can swing that, *if* it can
demonstrate that the "quotation" is not in fact copying.

As I wrote to another list, SCO has to prove several things to
demonstrate copyright infringement.  It's now focusing on contract
rights, but the case is similar:

  - Establish ownership of the work (currently under dispute).
  - Establish similarity of alleged infringing work.
  - Demonstrate, against defense arguments, a plausible line of
    derivation (copying isn't copying if you don't copy -- it's
    independent derivation).  This could be muddled if, say, the 
    allegedly infringed work is itself based on works which have been
    widely distributed with an explicit or implicit permission to copy
    and/or distribute.  The BSD/AT&T wars weigh against SCO here.
  - Demonstrate, against defense arguments, that the specific form of
    the work is not functional.
  - Demonstrate, against defense arguments, that the copy is infringing,
    rather than permitted under one of the enumerated exeptions to the
    exclusive rights of authors.

In most cases, establishing copyright, copying, and infringement is
relatively straightforward.  In SCO's case, ownership itself is in
dispute.

> 	Given the sheer size of the Linux kernel, even if you omit the
> vast amount of the code which exists in device drivers, what is the
> probability that you'll find some few hundred lines of code peppered
> about the thing that look like any other C project.  Allow for
> translation into kernel normal form.  

    $ cd /usr/src/linux; find . -type f -name \*.c | xargs \
	grep -A4 -B4 -E '([^a-z]sco[^a-z]|caldera|santa cruz operation)' |
	wc -l
    1305

(2.4.20 kernel).

I certainly hope SCO's talking about something more than that.


> 	Now imagine that you're comparing two posix operating systems.
> Who here expects that you *won't* find somewhere under a KLOC of
> matching code?  
> 
> 	I mean really, what the hell can you DO in "hundreds of lines"
> of C?  It's not like this is python or LISP, where you can write
> entire applications in a few hundred lines.  Are these "hundreds of
> lines" contiguous?

Point taken.  We'll see.

As Don Marti's aware, SCO's deadline to reveal source in two weeks (stated
of 14 May) has expired.

Peace.

-- 
Karsten M. Self <kmself@ix.netcom.com>        http://kmself.home.netcom.com/
 What Part of "Gestalt" don't you understand?
   At the sound of the toner, boycott Lexmark:  trade restraint via DMCA.
    http://news.com.com/2100-1023-979791.html



More information about the linux-elitists mailing list