Slashdot Mirror
SGI Compares Linux & System V Source Code
mrgoatCEO writes "It seems SGI has finished up their test comparing SCO's Unix System V code and that of the Linux Kernel, according to ITWorld. SGI found that any similarities between the systems (amounting to only about 200 lines of code) have been removed in Linux Kernel 2.4.22, and added that the similarities were 'trivial in amount.'" This follows moves by SCO to terminate SGI's Unix license.
Keep a watch on those stock prices and insider trades. Its not like they are even trying to hide it any more.
... and it seemed to identify some things. However, as the article says, there is much that could possibly be in the code that violates copyrights that couldn't be picked up by something such as that. Honestly, how do you compare something like that? Sure, you could have a group of real humans compare the code, but that will be subjective... certainly whoever compares it is going to be looking at things from their own point of view. The problem is, there are very few neutral people towards this issue that would know how to compare such a thing. (Arguably, there are just as few on SCO's side :-)
So, SGI has something we've all been wanting access to, essentially the System V Source. They run a check against the Linux kernal for infingement and only tell us a summary of their findings? Why can't they be more specific? Why can't they say lines 100-110 of module X in Y in the linux kernal came up infringing. They don't have to reveal any "sco owned" code, just what is already opened sourced on the Linux side. I'm I missing something?
"To lead the people, you must walk behind them"
First and foremost, it's good to be kept honest -- having other people's IP slipped into your codebase by well-meaning (or otherwise) people is a risk in all development projects. At least with OSS, a company can find the code and get it removed; someday, this will probably be used by a company acting in good faith (as opposed to SCO).
Second, it'll be nice to have the GPL tested out in court if for no other reason than the ability to point to it and say "it's been tested, it stands up". Given that a lot of important development in the next 10-15 years will be utilizing the GPL as Free-as-in-Speech products slowly displace proprietary fundementals like operating systems, it's important to the industry in general to work out any "bugs" in the GPL and get a prescident established sooner than later.
And hey, it'll be nice to see McBride brought up on those fraud charges, too -- you can't defraud your investors like this and expect to skate unless you're buddies with the guy in the White House, after all.
Every year during my review, I just pray the words "slashdot.org" aren't mentioned.
Copyright law allows for reasonable copying, e.g., a few pages from a book. So surely 200 lines of code out of millions would simply be laughed out of court?
Can someone please tell me what SCO really stands for? No I don't mean their morals, which may appear to some untrained observers as unprincipled extortion. I mean does anyone have any good expansion of the initials SCO?
"This comparison revealed a few examples of line-by-line copying, but did not determine whether the code was owned by SCO or in the public domain."
If Comparator were run against the appropriate *BSD (AT&T public source, right?) I wonder how many of offending script segments would be taken as not SCO's/SysV.
So SGI ran Eric's program that makes MD5 sums from a source file - why not release those sums so people can do independant analysis? (I can just imagine SCO's lawyers reving up ... "your honor it's a derived work", "we own those numbers" ...)
Now publish the questionable ~200 lines of code to replace, let us replace them and be done with it!!!
- The point of SGI's comparison was to search for any potential matches between Unix System V and any contributions that SGI made to the Linux kernel, not to vet the software for the entire community, Estes said in an interview. "We are not making any kind of representation at all about anybody else's contributed code," he said.
So SCO could still potentially be right in saying that IBM submitted a lot of Unix code into Linux.Curiously, because SGI has access to both codebases, and know how to run the Comparator, SGI probably has a good idea how many non-SGI violations are in the code (eg. the details may be complicated, but surely they can guess within an order of magnitude). So for now, it's IBM and SGI knowing and not telling, SCO knowing and saying as many ludicrous things as possible, and the rest of us left to speculate until this thing goes to court.
I think a lot of readers have missed the business strategies behind this situation.
Personally, I think its one of the most brilliant exit strategies ever conceived! I didn't say it wasn't messy or the pissy way to go, but this IS an exit strategy in the works.
Even before the lawsuits, SCO was facing two prospects - either continue to lose market share and become irrelevant, or make some noise and see what happens.
They were forced into the latter, simply because they have to make bang for bucks to the officers, employees and stockholders of the company.
FACT: SCO's IP is worth something.
Even if they lose their suit, they wil have driven up the value of their IP - SOMEONE will aquire them. IBM? Probably not. Someone will.
So SCO is terminating UNIX licenses left and right, but no one even notices. Can anyone point to any decrease in sales due to this with any affected vendor? Does this amount to the general opinion being set that SCO is now being ignored and good luck with enforcing any such lawsuit?
Is this meek whimper the final end of SCO?
Rule #1 -- Politics always trumps technology.
The reason SGI is important is that one of the two pieces of "infringing" code that they showed in a slide show came from SGI (the other one was immediately discounted as not being infringement and nobody ever seems to mention that one anymore).
They probably had no idea where the code came from or did not think it was going to matter. But it was identified and if it really was an infringement it identified an actual guilty party (somebody at SGI). It seems they have to at least make a show of suing actual infringers and thus have been forced to sue SGI.