[linux-elitists] Feedback Argument
Karsten M. Self
Thu May 15 23:25:54 PDT 2003
on Thu, May 15, 2003 at 03:58:51PM -0700, Mister Bad (firstname.lastname@example.org) wrote:
> So, I'm reading more stuff about the SCO doohickey, and I'm wondering
> about the feedback argument.
> This argument says that, OK, assume that some code from Unix got
> into Linux. But, because SCO was a Linux distributor, they waived any
> IP rights they had by redistributing the supposedly infringing
> software under the GPL.
> In other words:
> +----------------+ infringing
> | SCO Unix ------------------> Linux kernel
> | | GPL |
> | Linux <----------------------+
> | | | GPL
> | +---------------------> World
> In this scenario, the Unix code went out to the Linux kernel, came
> back in as GPL'd source, and went out as GPL'd source and binaries.
> It's obvious from the terms of the GPL that SCO grants Free use of
> Linux to World of any modifications it made to Linux. But I'm not sure
> I get why that would affect any infringing code that was in Linux when
> SCO got it.
> I guess the analogy would be picking the mailman's pocket and putting
> his wallet in a package that you then send to your friend. Since the
> mailman hands over the goods to you friend, he's lost his claim to the
> wallet. Seems fishy to me.
> Am I missing something here? Or does this argument lack merit?
Two related concepts.
First: to demonstrate copyright infringment, SCO needs to demonstrate
1. A work sufficiently similar in whole or part to the allegedly
2. A chain of derivation from the infringed work to the copied work.
That is, copyright protects works from copying. It doesn't protect
against independent derivation. Nor does copyright protect ideas,
concepts, or methods embodied in a work -- only the expression of
3. Proof that the copy is in fact infringing. This last bit is where
most copyright cases hinge. It's generally pretty easy to determine
that works are or are not similar, and to at least demonstrate a
plausible route from the original to the copy. The rest of the
question is whether or not there was a right to copy the work --
joint authorship, prior approval, exemptions to exclusive rights
(eg: fair use, or the other enumerated exceptions in 17 USC 107-122).
I suspect that this last part is where the claims made by SCO will fall.
We've seen a number of "intellectual property" claims by SCO, starting
with its 10-K filing which spoke of trademark infringement. The initial
buzz about a complaint was thought to center on patent until Don showed
there were none. As previously noted, ESR and I have largely demolished
any deep claims to trade secrecy (though limited scopes may remain).
The current thrust appears to be copyright claims.
Problem being, as Mr. Bad shows: SCO's been releasing the same source
under the GPL since 2001 in the form of Caldera OpenLinux.
SCO pleads that they were unaware of the issue "until the last couple of
years", and that they only got around to reviewing code recently.
That just doesn't wash with me. The time to perform due dilligence on
codebases was *prior* to the merger with Caldera. It's not as if SCO
can claim that they couldn't get their hands on Linux kernel source --
this is free software, freely available, and there for the taking on
archives around the world.
At the same time, SCO are taking an extremely stringent position
regarding IBM's obligations under a 17 year old licensing agreement,
covering code and technology thoroughly difused throughout the world.
Yet when it comes to SCO's promises under GPL (which its initial legal
complaint didn't even recognize as encompassing copyright protections),
SCO says, in effect, "wait, we didn't really mean that". SCO is talking
out both sides of its mouth regarding contract obligations.
There's an argument to be made that a company which finds its code has
been misappropriated, included in a free software project, and
distributed by third parties, would be justified in requesting removal
of that source and pursuing remedies against the person responsible for
the original misappropriation. Failure to perform due dilligence in
the course of a coporate merger, distributing the work(s) in question
itself for three years, then requesting an "undo", accompanied by a $1b
claim, seems completely disproportionate, if not wholly groundless. SCO
is in effect suing IBM for SCO's own incompetence in software
This is also an instance of the question which is posed frequently on
discussions of the GPL: "What happens if a programmer or company
decides to 'take back' their GPL'd code release?" The usual answer is
that, once an authorized release has been made, _that particular
version_ of the code can't be "taken back" based on a number of
arguments ranging from breach of contract to estoppel to fraud to
misrepresentation. SCO now finds itself in this situation. Trying to
undo its GPLd releases is likely to bring it yet more grief.
In your mailman analogy, it's more as if the mailman's wallet were
picked, then returned in a mix of packages, which the mailman was
obliged to examine closely to determine suitability for delivery.
*If* there was in fact misappropriation of SCO code in the first
instance, there is a cause for some complaint. SCO's failure to address
this in the first case, its inappropriate response in the second, and
its amazingly fluid legal complaint, don't help its argument.
Karsten M. Self <email@example.com> http://kmself.home.netcom.com/
What Part of "Gestalt" don't you understand?
"Charming man," he said. "I wish I had a daughter so I could forbid
her to marry one ..."
More information about the linux-elitists