Randall Davis: IBM Has No SCO Code 405
Mick Ohrberg writes "As reported by Groklaw, Randall Davis, renowned professor of Computer Science at MIT has after an extensive search found no evidence of SCO's claims that IBM has incorporated parts of the Unix System V code. Davis says "Accordingly, the IBM Code cannot be said, in my opinion, to be a modification or a derivative work based on the Unix System V Code." Surprised, anyone?"
Re:Finally... (Score:1, Informative)
Not quite..
An MIT professor isn't a court, although he could be used as an expert if IBM needs to. This is not that big of news imho. Other studies have already found this.
Thanks Professor Davis... and thanks ESR... (Score:5, Informative)
ESR [catb.org] deserves three cheers for 'scratching his itch', making a tool to compare copyrighted code. To have it actually used in the SCO case which was the annoying impetus for its creation (AFAICT) has to be a nice feeling.
I'm not an ESR fanboy, but I'll give him props when I think he deserves it and in this case I think he does.
--LP
Read the PDF... (Score:5, Informative)
a judge will weigh. (Score:5, Informative)
If I recall correctly, Randy told me that he has served as a special master in several cases.
Re:question (Score:5, Informative)
Re:Finally... (Score:5, Informative)
This has been gone over at length on Groklaw. IBM HAS taken action. No matter what SCO does, IBM still has a huge countersuit under something called Lanham Act [internet.com]. Methinks SCO is in a bit of trouble
Re:Finally... (Score:3, Informative)
Whether they still have any patents or copyrights on the functionality of UNIX remains to be seen,...
Ummm, no it doesn't. We already know that SCO doesn't have any patents, and there's no such thing as a copyright on functionality. We copyright code, not functionality.Re: a judge will weigh. (Score:5, Informative)
One would hope, as a matter of fact, the SCOexpert would be required to show where he found matches. That (if it exists) can be explicitly shown to the court.
Then, if need be, the experts can argue over wether or not they match.
Kind of like fingerprints. The suspect's fingerprints are entered as evidence, as are fingerprints found at the scene. The experts can then argue about wether or not they match. But until those fingerprints are presented and accepted as evidence, there is no weighing of testimony to be done.
Re:At $550 per hour... (Score:4, Informative)
I expect it took more than 10 hours for him to write that document and painstakingly verify its accuracy and wording to avoid perjury. He'll likely spend significant amounts of time testifying in person on the subject as well.
That rate is high but within reason for top-end expert witnesses (which is exactly what he is.) It's not uncommon for renowned professors to make a substantial second income by acting as an expert witness (very common in the chemistry and biology fields, at least.)
Finally, IBM would not even blink if they were handed a bill for several hundred hours at $550 each on this issue. They may even get some of the money back, depending on the details of the final settlement and the subsequent SCO bankruptcy.
Re:Read the PDF's methodology (Score:3, Informative)
1. ideas
2. purposes
3. procedures
4. processes
5. system
6. method of operations
7. facts
8. unoriginal elements
WOW! Okey Doke. So, now the IBM legal team is really looking for "copy-cat" aspect of which we, the community, are certain there aren't any (save for a few comments).
1. Ideas can't be copyrighted (but they can be patented).
2. Purposes can't be copyrighted.
3. Procedures can't be copyrighted (but they can be patented).
4. Processes can't be copyrighted (but they can be patented).
5. Systems can't be copyrighted (but they can be patented).
6. Methods of operation can't be copyrighted (but they can be patented).
7. Facts can't be copyrighted.
8. Unoriginal elements can't be copyrighted.
*Not* exponential Re:Scope (Score:2, Informative)
That's incorrect. Comparator makes one pass over all the lines of code, computes a hash values for line triplets, sorts that hash value, and then looks for matches in adjacent values of the sorted list.
It is not an n squared algorithm, it's n log n.
By my calculations, looking at 200 times as much SCO code (6M vs 27k) would take less than 3 times as long.
Mike
Re:question (Score:5, Informative)
Citing the GPL:
Clear enough?Re:Really??!! (Score:3, Informative)
To be fair, noone says they aren't, except an expert who was payed a lot of money (550/hr) by IBM to say so.
To you and I that would be a lot of money, but TSG is willing to pay BS&F many multiples of that (their law firm charges over $600/hr per lawyer, more for courtroom time. Mr. Silver was paid for 3 hours in this last hearing AND HE SLEPT THROUGH PART OF IT) for their expertice, why should IBM be less willing to spend money for expertice on their side?
As far as Mr. Silver, I would be willing to sleep for a few hours at $600 per hour, where do I sign up?
Randall Davis was paid a consulting fee (less than $6000 as far as I can determine)to give his expert opinion, not to say what IBM wanted. If what he found did not fit what IBM wanted, he would have still been paid, but his opinion would not have been used.
But of course, he's a lawyer, and they're well known for their morals and are among the most honorable men and women on the earth. They would never cheapen their personal values, nor lie and twist evidence, for cash.
But of course you are wrong and he IS NOT a lawyer, he is a professor at MIT and the developer of the standard test used by the 10th curcuit court - the Abstract/Filter/Compare test -which makes him a very big gun indeed, quite like smaching the SCO Group mosquito with a piledriver - massive overkill for such vermin.
Heh, paragraph 30 (Score:5, Informative)
I think that pretty much sums up this whole case from the beginning.
Re:Counter example would have helped. (Score:3, Informative)
That's the SCO argument, isn't it, that derivative works are theirs by copyright. Trouble is that the law is different. Copyright covers the representation not the idea. And for software having malloc() in two pieces of code doesn't rise to copyright violation.
Friendly logic lesson. (Score:3, Informative)
Your logic is flawed. If not true, it's unclear whether it's interesting.
Your conclusion would be true if your premise were "interesting if and only if true."
Re:Wait... (Score:3, Informative)
"And C++ programming languages, we own those" -Darl McBride
And nobody was more surprised to hear it than Bjarne Stroustrup [att.com] himself!
Duh! Can't you READ?! (Score:3, Informative)
Check.
Columbia Law Review article on "the Legal protection of Computer Programs". Check.
Software Law journal article on "The Nature of Software and its Consequences for Establishing and Evaluating Similarity. BIG Check.
Court Expert on Software Copyright Infringement. Check.
Retained by the DOJ to investigate copyright theft (and subsequent cover up) by the FBI, NSA, DEA, US Customs, and DIA. Check.
Served as chairman of the National Academy of Sciences study on ip rights and emerging information infrastructure. Check.
Retained as an expert in over 30 cases of ip infringement. Check.
Yeah, I don't care if you aren't impressed with his write up. He's got the skills to pay the bills.
He ran some already-written software, manually verified a handful of results, and reported what the software said. Anyone who can operate a computer can do that.
Actually, its the MANUALLY verified results that "anyone who can operate a computer" CAN'T do. The above C.V. gives gravitas to his methodology, choice of programs, modifications to said programs, and (most importantly) manual verification.
Re:What did Didio see? (Score:5, Informative)
Davis saw no -infringing- code. That's not the same thing as seeing no common code. Copying a ten-line function almost verbatim is likely a copyright violation. Ending up with a handful of lines that look a lot alike, by contrast, is often just an unavoidable side-effect of writing two pieces of code that do the same thing.
Small blocks of code (under 3 lines) are generally not protected by copyright (unless we're talking obfuscated C lines). Even larger blocks of code may not be protected, depending on content.
For example, in many cases (drivers come to mind), there may only be exactly one way to do something (e.g. you must set this register to 1, wait 5 ms, set it to zero, wait 5 ms, then set a second register to 1), in which case those specific bits are unlikely to be copyrightable at all, even if they represent a fairly substantial number of lines of code.
Also, in order for the code to be infringing, it must have been taken from AT&T UNIX or its descendants, and must not have been put there by someone who owned copyright on said code. That means that A. the code must not have been placed there by anyone working for SCO or Novell, and B. the code must not have come from a third, shared source such as BSD. A very large chunk of SCO's UNIX code fails the "B." test, and SCO was an active contributor to Linux, so many cases where code appears the same could easily fail the "A." test as well.
Re:Finally... (Score:3, Informative)
Yep. You can't copyright ideas, only a particular expression of ideas. In as much as the code -is- the idea, or implements a standard, it can't be copyrighted (e.g. interfaces). Patents cover ideas, not copyright.
Not that SCO hasn't argued exactly the opposite! They've been saying "UNIX concepts and methods" have been infringed as in the press, and even in the courts. It hasn't flown in court at all. The only examples they have of 'infringement' are exactly the kind of standard interfaces that can't be protected.
So basically their entire court case is based on a false reading of copyright law that, judging by the times on comments, would have taken them about 13 minutes of research to disprove. Pathetic.
Re:In other news... (Score:1, Informative)
So while it is not conclusive, it appears that rocks are probably solids.
Ice is also pretty soft, as rocks go. It only has a mohs hardness of about 2.5. (This is notable because people always imagine glacial ice carving up rock, but it doesn't, at least not directly. The ice is too soft. Glaciers erode, and very effective at doing so, by many mechanisms, but much of that erosion is accomplished by abrading the underlying rocks with pieces of non-ice rock frozen in to the bottom of the glacier, or by "plucking" wherein the glacier freezes around a bit of rock and yanks it out of the ground.)
Anyway, while ice may be rock, magma can sometimes qualify as wet. Depending on conditions and composition it usually contains a few % water, and some rare magmas may have up to 10 weight-percent dissolved water. Water aids in melting (both ice and other rocks), and is present to varying degrees in all magma. Mid-ocean ridge type magmas tend to be rather dry, but most crustal melts have a high fraction of water. The water content is very important to many magmatic processes: it is a key variable affecting the violence of eruptions if a body of magma reaches the surface; it affects intrusive metamorphism; and it allows rocks to melt that couldn't even approach melting temperature if they were dry.