SCO Defends Against Open Source Advocates

SCO defended itself against criticism by the open source community, saying Unix code used in Linux comes from its own, copyrighted version of Unix, not - as Linux advocates argued last week - earlier versions that have been released into open source.

SCO also said the General Public License (GPL), a popular license for releasing software into the open source community, violates U.S. and international copyright law.

SCO, which owns intellectual property of Unix, is fighting a legal and marketing campaign to show that Linux contains sufficient proprietary Unix code that Linux distributors and users are ripping off SCO. SCO sued IBM in March, claiming that IBM included proprietary Unix source code in Linux, and later SCO warned Linux users that they, too, could be subject to intellectual property lawsuits if they failed to obtain legitimate licenses from SCO. SCO introduced a $699 license this month for Linux users.

SCO has been closely guarding its evidence - the allegedly stolen source code - disclosing it only people willing to sign a nondisclosure agreement. However, last week SCO disclosed some of the code at a presentation at its SCO Forum conference in Las Vegas. Linux advocates, including Bruce Perens and Eric Raymond, obtained copies of SCO's slides, and posted responses to the Web, saying that SCO's own evidence undercut its case.

Perens and Raymond said that some of the code disclosed on the slides, governing memory allocation, comes from early, "ancestral" versions of Unix that were released into open source by SCO itself, while SCO was doing business as Caldera.

But SCO said that, while ancestral Unix versions have earlier versions of the code, the code was refined in SVR4.1, and it's the later version of the code - still proprietary to SCO - that appears in Linux.

Chris Sontag, senior vice president and general manager of SCOsource, the company's business unit for licensing SCO's intellectual property, said another vendor copied proprietary memory allocation code from Unix into Linux, removed the original copyright notices and attached its own. SCO would not identify which company did the label-switch.

Linux advocates also said last week that the memory allocation code is only used in versions of Linux for IA-64 systems, such as systems running on the Intel Itanium chip, representing fewer than 10 percent of Linux installations. The vast majority of Linux installations are running on IA-32 systems such as Intel's Xeon processors.

"What's at issue is that there is copyrighted Unix System V code, Version 4.1 code, copied into Linux. Whether it is used broadly or not, it is widely published and available. SCO's copyright is stripped out and others are taking credit in violation of the copyright," Sontag said.

Perens said that one of the examples of allegedly stolen code shown by SCO last week was, in fact, from Berkeley Packet Filter (BPF) routines taken from BSD, which is covered by an open source license.

But Sontag said the BPF routines were not intended to be an example of stolen code, but rather a demonstration of how SCO was able to detect "obfuscated" code, or code that had been altered slightly to disguise its origins. The slide displaying the code should have been written differently to reflect that intention, he said.

"It was an example of our ability to find moderately changed or obfuscated code, it was not an example we are using in court," Sontag said. "If they want to go off and make a big defense on that, they are welcome to it."

Sontag said the code examples SCO chose to disclose last week were not its best examples, merely the most easily understood ones. Perens had said the examples SCO disclosed were likely to be SCO's best, and underscored the weakness of SCO's case.

"He's wrong, he doesn't have examples of the evidence. We do. He is trying to put a happy face on a problematic situation for the Linux community," Sontag said. "Try as they might to come up with arguments to bolster their position, the facts and everything we know are extremely strong in SCO's favor."

SCO also focused criticism on the GPL, which is the license for many open source projects, including Linux. The license states that GPLed software and source code must be available to anyone. Modifications to GPL software are subject to the same provisions.

The Free Software Foundation developed the GPL, and defines free software on its Web site.

Linux advocates say that SCO undercut its own case by releasing its own version of Linux under the GPL. The SCO version of Linux contains the disputed code and - even if the code was once proprietary - SCO released it into open source when it released its own Linux, the advocates argue.

However, Sontag said that argument holds no water because SCO never intended to release its proprietary code into open source. "U.S. and international copyright law asserts you cannot inadvertently and accidently assign your copyright to someone else," Sontag said.

Moreover, SCO said its proprietary code in Linux does not meet the definition of free software as stated in the Linux GPL.

"The Linux GPL itself asserts that the valid legal copyright holder has to place a notice at the beginning of their copyrighted work, the source code, identifying the code and the GPL. It requires an overt action. SCO has not contributed its code, and as soon as we became aware of the copyright violation we suspended our distribution," Sontag said.

Certainly not, but they did purchase most rights to the code so it is theirs whether they actually developed it or not

That's what they say, but the person they bought it from (Novell) insists otherwise. The problem with insisting that they bought the rights to this code is that the rights to the code had been transferred to Berkeley (and the BSD group) by AT&T before AT&T sold the remaining Unix intellectual property to Novell. Just ask Novell or read the Novell contract; they have been trying to tell SCO this for quite a while, but SCO won't listen.

Your argument seems equivalent to claiming one can't actually own their house unless they were the ones who originally constructed it.

Not at all. My argument is that you don't own every house built with a certain floorplan just because you bought a house built from those floorplans.

Using your analogy, SCO purchased a house from real estate agent named Novell. Novell had purchased that house from a builder named AT&T, who originally designed the house but gave the rights to the plans to a non-profit architectual firm named Berkeley, who let anyone use the plans who wanted to do so, as long as they didn't charge anyone else for the plans.

Now, SCO comes along and not only wants to sell houses based on the design that legally belong to Berkeley, but are claiming that they are the ones who designed the house in the first place, and that all the other people who used Berkeley's "free" design owe them royalties, because they are the rightful owners of the design (based only upon the fact that they bought a house using that design from the Novell realtor).

The fact is, SCO didn't write this code. They also didn't purchase this code (they claim otherwise, but Novell insists that all they purchased was the name "Unix", and in any case Novell couldn't sell them what Novell doesn't own either. And Novell asserts that it didn't own this code).

Golden Eagle wrote:
Because it is a "trade secret", that they wish to keep secret?

It might have been a trade secret at one time, but it is no longer a trade secret. Anything in Linux (or any Open Source project) is not a secret by definition.

By the way, some of the alleged "trade secrets" are so secret that SCO didn't write the code for them (JFS, NUMA scheduler, RCU, etc.). SCO will never prove that code written by IBM and Sequent is a SCO trade secret. That's a big part of the SCO v. IBM case, and they will lose that part of the case.

Golden Eagle wrote:
Here's an analogy - admittedly of signifcantly more importance - but what if classified US Government material got into a book of some sort, would you then expect the government to publicly say "It starts on page 87 and ends on page 95"? I sure hope not.

No, but that's a very flawed analogy.

Saying that did happen, though. Would you then say that the US government could go to everyone who bought the book and say, "Pay us $1,000 per copy for that book and let us lock it in this sealed lexan box so you can see that you have it, but you can't read it. If you don't like that, you can turn the book over to us and never read it again." Would that be alright by you?

SCO is basically asking business users of Linux to give up their rights to use Linux source code as part of their "SCO license for Linux IP" or whatever they call their fraudulent "IP License."

Golden Eagle wrote:
Certainly not, but they did purchase most rights to the code so it is theirs whether they actually developed it or not. Your argument seems equivalent to claiming one can't actually own their house unless they were the ones who originally constructed it.

Actually, Caldera (now the SCO Group) purchased the rights to "UNIX" from Santa Cruze Operation (now Tarantella), who purchased those rights from Novell, who purchased those rights from AT&T's UNIX Systems Labs division. Caldera (now the SCO Group) also purchased all of the source code for UNIX, and all of the source code, trade names and development of the SCO UNIXware product, which is a UNIX version for Intel computers.

SCO did not purchase, just as an example, the copyrights and source code to IBM's Journaling File System (JFS). IBM developed the Journaling File System first for OS/2, then ported it to AIX (which is IBM's licensed UNIX product). Now, IBM has also ported JFS to Linux and released it open source as part of the Linux kernel. SCO claims that JFS is one of SCO's "trade secrets" and that IBM doesn't really own it because of a very tortured interpretation of the "derivative works" clause in the AT&T/IBM license agreement for UNIX. So, even though SCO didn't write it and doesn't claim a copyright on JFS, SCO claims that IBM can't sell JFS on any platform other than UNIX (AIX), and that IBM can't release JFS as open source or port it to Linux. The judge and jury haven't ruled on this, but I doubt they will be impressed by SCO's claims.

SCO is making similar claims about the NUMA schedulare and RCU code (which was written entirely at Sequent, now a subsidiary of IBM).

Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.