Responding to doubts about its claim that its Unix source code was illegally incorporated into Linux by IBM, SCO put some of the actual code in question on display for Unix conference attendees in Las Vegas this week.
Lines of the code, which SCO displayed during a slide presentation on its suit against IBM, apparently have served both to bolster the Lindon, Utah-based company’s complaints about copyright infringement and to cast doubt on the company’s ownership claims. Many of the SCO-hosted Unix forum attendees who viewed the code reported that, indeed, large pieces of the code — including programmer comments and spelling errors — seem to have been copied directly into Linux.
On the other hand, the code SCO presented in its slideshow comparison was reportedly checked by technology experts who dated some of the code as far back as the early 1970s, a revelation that could cast doubt on SCO’s claims that its copyrighted intellectual property – System V Unix source code – was improperly used in Linux.
Yankee Group senior analyst Dana Gardner told TechNewsWorld that SCO’s code revelations are simply part of a public-relations campaign that mirrors the software company’s legal campaign, which might take years to resolve.
Gardner said SCO is showing enough code to pique the interest of the industry but not enough to allow a remedy to the alleged copyright infringements. Some have said it would be relatively simple to dispense with the SCO code or any of the modules that use it if the industry knew which lines were in violation.
“It seems as though they’re trying to have it both ways,” he said. “They’re claiming that they’re being injured to the tune of $3 billion plus, but they can’t show you where it hurts.”
Need for Uncertainty
One of SCO’s biggest burdens is to prove that IBM or anyone else using Linux has copied directly from SCO’s Unix code — as opposed to reproducing software functionality, which is an acceptable practice, software legal expert Phil Albert told TechNewsWorld.
Albert, a partner with intellectual property law firm Townsend and Townsend and Crew, said SCO will face a challenge when it must get specific about its source-code claims because programmers will be able to write out or around the disputed portions of code.
“That’s a problem for SCO,” he told TechNewsWorld. “They need the uncertainty. If they are too specific, the problem will get fixed.”
In response, Chris Sontag, SCO’s senior vice president and general manager of SCO Source, told TechNewsWorld that the company has now unveiled the offending code and that it can be remedied.
“The vast majority of the code [in violation] is the derivative work from IBM, so that’s a great place to start,” Sontag stated. “We’re talking about more than one million lines of code that can be remedied.”
Unix of Yore
Albert said that regardless of how old the code in question is, SCO does hold the copyright that goes back to the origins of Unix. The company has been arguing that the case hinges mainly on whether Linux is a derivative of Unix.
“Unix is an old operating system, so it’s been around a while, but that doesn’t necessarily mean SCO has less of an issue with it,” said Albert. “The argument is not that SCO created these portions; even if they didn’t, they still own the copyright on it.”
SCO’s Sontag rejected claims that the copied code dates back to the origins of Unix, citing SCO’s several-year-old System V version 4.1 as the main version of software where illegal copying occurred.
Gardner said that although the “componentization” and borrowing of software code has been viewed as a good and necessary thing in the industry, the line that divides illegal from legal behavior is fuzzy.
Slower Than Software
Gardner argued that it is unrealistic to expect legal oversight to keep up with the process of software development.
“It’s almost nonsensical to think the legal domain could keep up,” he said. “To think that there should be a level of oversight to something so fast-moving, creative and dynamic is pie-in-the-sky thinking.” To hold software activity hostage by a series of legal maneuvers “doesn’t seem possible,” he added.
Although the case is not expected to begin making its way through court until next year, the presentation of code, efforts to limit or expel discovery material and other legal maneuvers dealing with the code will be key, according to Jeff Berkowitz, a partner with intellectual law firm Finnegan-Henderson.
“The litigation process in this case is going to be significant,” Berkowitz told TechNewsWorld.
He added that the damages stage of the case, which could take two years or longer with appeals, also will be significant because SCO might pursue a theory based not only on IBM’s software sales, but also on Big Blue’s associated hardware sales, which could raise the stakes even higher than the current claim of several billion dollars.
It was a long time since I’ve read something so misguided and misinformed.
Do some research before you post a story, will you? If you did such a thing as a 10-minute research before posting this story it could come up much better (Hint -use google news service – type SCO in that little box – you’d be surprised at the results!)
SCO’s so-called "evidence" originated from code from 1973 that was TWICE released under BSD license by the rightful copywrite holder since.
It does matter how old he code it, because according to ruling from 1993, AT&T holds no exclusive rights on any code in the BSD linux.
The "evidence" SCO published is just another proof their claims hold no merit. I hope you’ll do a better job at creating a more balanced report next time.
Let me say first that I sincerely hope SCO fails in its efforts. However …
If there turns out to be infringing copyrighted code in Linux, and the customer _makes copies_ of Linux for herself, for her business or for others, then the customer also is liable for infringement claims.
Please see these articles before accepting the knee-jerk attitude that "if any code in Linux is the same as code in SCO, then there must be infringement".
Linus Torvalds’ analysis of SCO presentation: http://www.eweek.com/article2/0,3959,1227081,00.asp
And Bruce Perens: http://www.perens.com/SCO/SCOSlideShow.html
Bottom line is, as Linus says, "They are smoking crack".
SCO is purposely confusing the amount of code they claim to be a copyright infringement with what they claim to be unauthorized derivative works.
When they throw numbers around they combine the two and hope the media will assume they are referring to copyright infringement numbers.
The very shaky derivative works arguments worst case are a contract violation. There is no way possible they can claim copyright on "derivative" works. The legal dispute about derivative works is between SCO and the companies they have contracts with. That in no way strengthens their claim re: linux users.
Next time McBride or Sontag offer an interview on their FUD tour, ask them to clarify these numbers.
How many lines of code/files in Linux do you claim infringe Unix copyrights?
How many lines of code/files in Linux do you claim are derivative works?
Do you claim to own the copyright on the derivative works?
Isn’t it true the the only legal standing you have with companies that sell Linux but have no contract with SCO is on the basis of the copyright infringement claims?
Isn’t it true that there is no legal claim against the customers of linux sellers, if there is copyright infringement it is not the end-user committing it?
If you claim that you are suffering damages due to copyright infringement, aren’t you obligated to make all parties aware of your claim to allow them to mitigate the damages?
The first code snippet is an implementation of malloc() dating back to the ’70s. It was released under the BSD license twice, including by Caldera itself (the company that is now called SCO) in 2002 as part of the "Ancient Unix" collection.
The second code snippet, referring to code that has been allegedly obfuscated to conceal its origin, is not even SCO’s property. The code in SYSV was originally written at Berkeley as part of BSD. This code was later incorporated into AT&T UNIX (years before SCO bought SYSV), but the copyright attribution was removed and now SCO thinks it’s theirs.
What’s more, the code in Linux is *not* a deliberately obfuscated version of the original. It’s a clean-room rewrite of the same algorithm. It only looks similar to the BSD version because it follows the same specification (the spec even gives field names).