[plug] SCO and rewriting the kernel

[Leon Brooks]

On Sun, 8 Jun 2003 09:30, Chris Caston wrote:
> I'm wondering if it is real how long is it going to take Linux
> Developers to rewrite the kernel?

Um, what?

We'd be rewriting less than a thousand lines (if it were a thousand or 
more, SCO would be calling it "thousands of lines of code"). AFAICT - 
and SCO have been working hard to keep this difficult - the problem 
consists of a few sections of maybe a few score or few hundred lines, 
plus numerous "10 to 15 line" chunks which would easily fall under the 
birthday paradox and so wouldn't need fixing (I think they'd be fixed 
anyway).

I can't imagine that taking as long as a day, including testing, and 
expect distributor updates (except maybe SCO) to be pretty much 
complete within a week.

Note that this is worst case; following is what I had to say to the 
editor at ZDNet about it:

<quote>
SCO needs the publicity to pump up its stock prices and generally look 
worth buying. Even if, very much against the odds, they show that IBM 
did something naughty, that's between them and IBM only - and if IBM 
felt threatened to the tune of a gigabuck, I imagine that they have 
enough IP of their own to turn SCO into a smoking memory within weeks. 
SCO have absolutely no claim against anyone else, move along folks, 
ignore the desperately noisy fellow here, nothing to see.

Their new-found copyright amendment is next to useless for them since 
the alleged copyright transfer wasn't registered with the Copyright 
Office (ie no damages to be had on this point, the worst they can do is 
block further transgressions and they still have to closely identify 
the transgressions to do that; Novell has only ever registered two 
copyright transfers, one to Corel, none to The Santa Cruz Operation, 
The SCO Group, Caldera, Tarantella or anyone related).

They've also alienated a big slice of their own customer base. I, for 
one, have stopped installing or repairing SCO systems. I've migrated 
all but one of my own SCO-based customers away to Mandrake Linux, and 
that last one is scheduled for conversion in a few weeks. No matter how 
much brown-nosing they do later, I won't ever recommend or support a 
SCO system again.

As to the "against the odds", they really have made things very hard for 
themselves. Their FTP site still even today offers free copies of the 
source code that they claim is at risk, they haven't identified the 
code anything like clearly enough for anyone to minimize any damage 
(their claim that Linux advocates will clean up the code and magic away 
tens to hundreds of thousands of source CDs and archives worldwide is 
risible), Ransom Love went on a "unification" drive and SCO's website 
lists some Caldera contributions to the Linux kernel. They've made no 
mention of common sources like ancient Unix (which they themselves 
open-sourced), textbooks and the like. The onus is very heavily on them 
to prove that they do own the code in question exclusively, that it 
wasn't developed independently/coincidentally, that the same code was 
knowingly inserted by IBM and not somebody else (including themselves), 
and so on. That's a tough row to hoe even for a well organised company, 
and SCO keep changing their story.
</quote>

Cheers; Leon