Fog Creek Software
Discussion Board

Analyst reports on SCO/Linux code comparison

S. Tanna
Thursday, June 5, 2003

Doesn't this just contribute to the Linux bad boy image, and make us want to use it that much more? ;-P

Co.'  harhar.

Brian R.
Friday, June 6, 2003

Actually this article brought me an idea I'd never noticed before.

It's a double-whammy:
1. Cast legal doubt over Linux, and scare the decision makers away from it. This has the added benefit of casting that same doubt over any open-source product. Who knows where those Apache bearded geeks got *their* code?
2. Hit IBM. Make the "Blues Boys" spend resources trying to calm down their customers.

My vote goes for the "fiendish-MS-plot":)
"Suravye ninto manshima taishite (Peace favor your sword)" (Shienaran salute)
"Life is a dream from which we all must wake before we can dream again" (Amys, Aiel Wise One)

Paulo Caetano
Friday, June 6, 2003

I'm still curious as to which bit of the kernel is actually alledged to be infringing.

It can't be the what I would regard as core kernel functionality, e.g. scheduling and process control, VM, etc. because all this has been written and rewritten by non-IBM core kernel developers.

Similarly, it can't be any of the peripheral device drivers, simply because Sys V and Linux are seperated by so many years of obsolesence.

Can anyone give some reasonable guesses as to what similar between Sys V and Linux, that goes beyond generic implementations of public domain algorithms?

Chris Davies
Friday, June 6, 2003

Chris Dibona was on The Screen Savers talking about this issue. Linus named him as one of the people he would trust to check and see whether there's infringement. However, Chris stated he wouldn't do it, because the NDA that SCO is requiring would basically give him the ability to say "yes" or "no", and that's it. He wouldn't even be able to tell the world WHAT code was being infringed.

So, don't expect anything more than the type of article you just read, at least until it gets into court somehow (or someone convinces SCO to lighten up on the NDA).

Brad Wilson (
Friday, June 6, 2003

Finding duplication is not enough.  SCO could have taken Linux code and put it into their Unix.

T. Norman
Friday, June 6, 2003

"The month of June is show-and-tell time," McBride said. "Everybody's been clamoring for the code...and we're going to show hundreds of lines of code."

Hundreds of lines of code?  Its not uncommon to write several hundred lines of code per day - or more.  What is the rule of thumb, 10 lines of code per day average?  That includes the entire life cycle of a project: design, writing, debugging, integration, field support, ship party, etc.  So, SCO's claim is about a couple of engineer-months worth of work, max?  Once discovery takes place, this will be over in days.

As for how the code got into Linux, could SCO have done it themselves?  Cringely makes a convincing case for it:

"Open Unix 8 is the first step in implementing the vision of the pending new company," said Ransom Love, president and CEO of Caldera Systems in a company press release way back when.  "It combines the heritage of Unix with the momentum of Linux, and will be our premiere product for data intensive applications like database, email and supply chain management. The incorporation of the Linux application engine into the UnixWare kernel essentially redefines the direction of the product, and motivates a new brand identity -- Open Unix."

"When we talk about unifying Unix and Linux, the two have a huge amount in common. A lot of people are running their businesses on Unix, while Linux has a tremendous population on Web servers and other front-end servers. So we are taking both and combining them into one platform."

Nat Ersoz
Friday, June 6, 2003

"Finding duplication is not enough.  SCO could have taken Linux code and put it into their Unix."

This is the basic argument.  Look for article links at that talk about why they believe this is the case.

Brian R.
Friday, June 6, 2003

Having common code, is in my opinion, not sufficient to prove SCO's allegation in itself - but it is consistent with SCO's allegation (and a number of possible alternatives).  Frankly we haven't been given enough info to make a judgement.

I don't really personally care one way or the other (don't use Linux), but I am interested in this story.

A. We don't know if it's .h's or .c's (or I've seen nothing to indicate either way).  If it is in a .h (or I guess some .c's) it could be stuff like function/structure definitions conforming to some outside spec (like Posix, ISO C headers, some hardware spec, etc.)

B. Personally I don't think it's enough even if there is common actual code between products A and B

1. Code could be copied from A into B

2. Code could be copied from B into A

3. And few people seem to be discussing this one, but code could originate in a common ancestors C, D, E, F, etc and put into both A and B.  Such common ancestors might include public domain code, code under BSD license, code published by hardware manufacturers (Intel, device drivers by manufacturers etc.), possibly from books or whatever

S. Tanna
Friday, June 6, 2003

*  Recent Topics

*  Fog Creek Home