Slashdot Mirror
Latest SCO News
SCO has discovered an amendment to their contract with Novell that may clarify that they did purchase the copyright to System V after all. Heise has an interview in German with a former employee. Cringely says SCO probably was responsible for any duplicated code itself, with a theory that is quite plausible. One non-programmer corporate analyst has looked at SCO's alleged evidence. And SCO has another press conference today.
Big fucking deal. What's a non-programmer going to say about code? That's like going to a farmer and asking him how to fly an airplane, or a pilot, and asking him when the best time is to plant the corn.
From the article:
So Novell agrees that the information in this ammendment appears to be legit, but they can't verify that this ammendment actually occurred because they don't have a copy of it themselves?
Could it be that SCO happened to "create" this ammendment and then convienently "find it in a filing cabinet" ?
In C++, friends can touch each others private parts.
how can you patent 1s and 0s
On the other hand, how can you pattent a bunch of atoms?
Don't take me wrong, I agree with you, but not on the basis of 1s and 0s.
Of course, this is also a problem with US copyright law. Copyrights are so nebulous and easily transferred that it's almost impossible for end-users to keep track of whose IP they may be using. Registration with the US copyright office should be a requirement, not an option.
Not having the benefit of seeing the code I'll have to assumme these comments are fairly overwhelming evidence wise.
/* Beware all ye who enter here */? What if both System V and Linux stole the code from BSD? Was a SCO lacky sitting over the analyst's shoulder pointing out the high points? What if the "analyst" was a SCO plant from the beginning?
Thats a poor assumption. Whow knows what the context of the code is? Could this "non programmer analyst" determine if the code was really similar? What if the System V code came from the IDE subsystem, and the Linux code came from the networking subsystem? What if the comment was
There are just too many questions and not enough answers. I want to see the professional opinion of a kernel expert, and *then* I want to review it myself before I will start to agree that SCO might have a case. Until then, we have no idea what deals have gone on behind the scenes. There is just way too much money in play here for us not to be cynical.
Do you have Linux and a DotPal? Click here now!
Big f***ing deal. What's a non-programmer going to say about code? That's like going to a farmer and asking him how to fly an airplane, or a pilot, and asking him when the best time is to plant the corn.
I'd say it's more like asking an airplane pilot if two cows look the same.
Even more like asking a farmer if two planes look the same. They're designed to perform the same function (fly) with basically the same equipment (wings) so there will be some similarities and a lot of the differences will be in the technical details (the camber of the wing and the hydraulics that move the control surfaces) that an observer not fluent in airplane design might miss.
Granted, if the comments in question all contain the programmer's initials (something not uncommon where I work), then that would be pretty damning.
Could it be that SCO happened to "create" this ammendment and then convienently "find it in a filing cabinet" ?
Sure. And when found to be a falsified document then Novell would sue the everliving crap out of them. Fraud, forgery, harm to business, and probably a dozen or so other civil and criminal charges would be filed.
Even I don't think SCO is that stupid.
In any case, unless the ammendment was filed with the Copyright office it makes very little difference -- all SCO could do is sue Linux companies and users to cease further infringement, not monetary damages. They couldn't even recoup legal costs for the cases.
I should be able to patent the methods I used to achieve that, as well.
IF (and ONLY if) you (and for this discussion, "you" is any developer, not avalys in particular) actually innovated to do so. Changing from one well-known algorithm or data structure to another well-known one with better average or worst-case performance IS NOT INNOVATION. It's what you (hopefully) learned to do in school.
Yes, if you really, TRULY, come up with some new algorithm that no-one thought of before, yes, you should be *allowed* to patent it.
But there are two very good reasons why you should think long and hard before you do:
1) Chances are, you *didn't* invent it -- you probably just independently arrived at a solution that HAS been done before. So, there's a considerable risk that it's *already* been patented somewhere else (e.g. IBM or Microsoft), or that there's prior art that's clearly NOT patent-encumbered.
2) More importantly to software engineering as a field of practice, a great many true innovators make it a point NOT to protect their innovation, but instead to share it with their collegues, with students, with anyone who's interested.
Yes, that may be bad for business for the short term, but as a field that's still very research oriented, it's better for everyone in the long term.
Imagine where we'd be today if Dijkstra (holy crap, I spelled that right on the first try?) had patented his shortest-path algorithm? If various process-scheduling algorithms were patented, instead of published in textbooks?
If you only care about the here and now and your back pocket, sure, patent a method -- if you can truly convince yourself that you've innovated, and were the FIRST to do so.
But if you're interested in furthering the field, or if you know full well that what you did is neat but not truly new, *innovative*, and *non-obvious* (even after the fact), consider sharing, and letting others build on your work, instead.
Xentax
You shouldn't verb words.
Consider errno.h (or its various "component" (i.e. "#include"-ed files).
Most non-programmers comparing instances of this file from different sources would think "hey, these are almost exact copies!".
On the other hand, most programmers would be quite aware that they almost have to be exact copies: you need the "#define", you need the error name, and you need the error value, and they need to be the same. You could even imagine the comments being the same, or at least very similar. Most programmers would understand that these values are needed for compliance with published POSIX-type standards. Non-programmers would not.
Come to think of it, maybe it's something silly like this that SCO is complaining about.
I haven't posted yet about this story because I couldn't figure out exactly what I wanted to say. I just knew that it made me sick every time I saw SCO vs. Linux or any time I thought about the lawsuit(s) involved.
I think it's a deeper, more disturbed form of the same sensation that I get when we discuss intellectual property laws on Slashdot.
What it all comes down to is this: I don't care if there are six or ten minor chunks of SCO code in Linux. I don't care where the Linux code came from. SCO is not good for humanity; SCO is a product. Linux, on the other hand, is good for humanity on a fundamental level.
It brings computing services to people across the world who otherwise couldn't afford it or who otherwise would be sending money to multi-billionaire Bill Gates instead of buying food. Thanks to Linux, these people can spend their money on real things that they need, while still participating in the global exchange of ideas and perhaps getting a toe-hold in the "modern" western world that.
Linux provides nonprofits like churches and community centers the ability to provide 'net access and document services to underpriveleged communities who otherwise wouldn't have access to these things or would have to depend on meager government funding to buy licenses (again while lining the pockets of the rich).
Linux provides some of the best hands-on education to young aspiring programmers and scientists that can be found anywhere, and it does so using best-of-breed tools, and at no charge.
Linux has fostered a community of international understanding between research organizations, governments, communities and even small groups of programmers and individuals. There are no borders in Linux, only individuals working together, smiling and one another and breaking barrier after barrier together.
Some look at this list and say "hmm... makes national borders irrelevent... helps the poor and not the rich... does not pay for labor in currency, but in the rewards of the product itself... is not strongly managed from the top down by anyone with fiscal authority, but is instead contributed to by a vast egalitarian labor pool..." and then they call Linux a form of communism and say that it needs to be eliminated, or at least that it is unethical to support Linux instead of for-profit companies.
Are these people insane? Linux is a boon to humanity. Anyone who can't see that is blind. If Linux is communism, it's time to take another look at communism, because it looks to me like a beautiful thing.
And in the meantime, I don't care where the code of Linux comes from. The fact is, SCO's never done thing one to help the human race, here they are busily exploiting it, rich and poor, young and old a like, just like so many other companies out there greedily trying to harm all our lives on the basis of "IP", and now they seem to think they can kill Linux off in the interest of making a buck, and that this would be a good, "moral" thing... and many mainstream analysts seem to make the basic assumption that if Linux did "steal" SCO's code, then this is indeed the case. Well, I disagree. Linux helps people at no charge. SCO makes a profit by exploiting people.
It's just wrong. It's backward. It makes me sick. I don't care where Linux's code comes from. SCO is no Linux and never will be, and if we end up with a nice, profitable SCO and a damaged (or defunct) Linux movement, I think the world will be a much worse place, not a much better one, no matter how much people harp about the "right" thing to do or the "rights" of copyright holders.
I almost feel like some kind of neo-flower-child. Screw the establishment. Help the people. If that's communism or if that's bad for business, so fscking what?
It's not wise to get in a bind over the comments of one analyst. Remember, analysts of this same sort said .com's would continue their rise. It may very well be in this analyst's interest to assert similarities in commenting style. Interestingly enough, there was no talk in the article about actual code being reproduced, only comments.
SCO's claim was that source code had been cut directly from "their" Unix code and added to Linux. This does not preclude someone from working on both projects. All this analyst's statements show is that the same people may have worked on both systems. This doesn't show a wholesale heist of intellectual property.
SCO may be taking a page from the M$ playbook by intentially pushing "evidence from experts" to the public. Will this analysts comments be mailed to Linux users? It wouldn't surprise me in the least. If SCO can drive FUD to the hearts of corporate types, they can all but force an IBM buyout.
This is great stuff for tech geeks, but publications that your boss is reading such as this article over as business week are what your boss(you know, the guy who pays your salary) are reading. I would say this whole debacle is having quite the intended effect.
mp3's are only for those with bad memories
My thought on all this is that that won't be known until such time as the code in question is actually released. At which point the Linux kernel admins will look at their records, hunt down the people who submitted those patches, and *ask them*, and probably call them as witnesses.
At this point things will probably get interesting, as there are open, public records that show bit by bit in pretty close detail every step of the development of the linux kernel, and who submitted what, and when, and why it was accepted-- and these records have been publicly available on the internet for years, and archives exist in various places, which would make these records impossible to change after the fact. SCO, meanwhile, if they have records at all of when and by who code was added to their materials, has no particular proof that those records are real and not faked (either by their lawyers or a malicious employee years previous trying to pass off open source code as his own work). Just a thought..
Of course had SCO simply begun all this by publicly saying "hey, these parts of the linux kernel are copied from code we own the copyright to", the linux kernel admins would have just about certainly simply checked out those sections of code and the people who wrote them, and, if there was an apparent infringement, removed and replaced the offending sections. I'm really really hoping that this fact will not escape the judge at trial.
(I'm assuming the linux kernel people wouldn't do something stupid like allow an anonymous patch into the kernel.)
Irritable, left-wing and possibly humorous bumper stickers and t-shirts
Interresting. I suppose SCO is doing a good job. Before we were arguing this whole thing is BS, and they side stepped our arguments.
/.
The steps;
1. SCO is off their rocker and their is likely no stolen code.
2. Novell interjects: Not only that but SCO does not even have the copyright to the code they claim is stolen.
3. SCO rebuts: Yes we do.
4. Well IBM should settle since the code is obviously stolen.
I do not know what 2 and 3 have to do with getting from 1 to 4, but this is the strange path which seems to have been taken on
I can see many ways that SCO could be cheating on this 'under NDA' information.
1. They modified the code (either Linux or SCO base) before showing it to the examiner, to make it look like it was closer than it actually was. Did the examiner bring his own copy of the source trees, or use those supplied by SCO?
2. The code was copied FROM Linux into SCO. Judicious back dating would be used to try to hide this fact.
3. A non-programmer doesn't understand code, and is probably only looking at the comments. In a million lines of code how hard is it to find similiar comments in unrelated sections of code.
4. Both the SCO and Linux programmers read the same books/articles, and the comments are based on what was there, thus giving similiar comments. Do the comments match up to Knuth's books?
5. Both SCO and Linux got the code from the same source. How much of BSD has SCO copied into their kernel?
6. The code in question is similiar to what you do in your beginning programming classes, and most programmers use similiar comments for that kind of thing. How many different ways are there to comment a bubble sort?
7. Someone within SCO supplied the code to Linux. Maybe in preperation for this case.
8. etc.
Until SCO allows someone capable of researching the origins of the code in question, I'll continue to believe that it is SCO that is in the wrong.
The code for both systems is already available to many people. Allowing others to see what they are complaining about won't make the suspect code disappear. If they just pointed at a bit of Linux code, they wouldn't even have to show their own code.
The only reason for not disclosing it before the trial is to gain time to hide their trail, or to deny IBM time to research their wild claims.
Who would win this election: Andrew Weiner vs Andrew Weiner's weiner.
Apparently the most telling evidence is that parts of the SCO code and Linux code include identical annotations made by developers when they wrote the programs, says DiDio, who compares such notes to the signature or fingerprint of a developer's work. "The fact that these appear to be transposed from Unix System V into Linux I find to be very damaging."
Now, who in the hell at IBM would be stupid enough to include identical source code 'annotations' (which I'm assuming are comments) when stealing code?
Similar annotations, maybe, but come on - even Joe Dumbass wouldn't be so fucking stupid so as to copy-and-paste proprietary source code from one to the other, comments and all.
Further, does SCO have proof that the infringing Linux code is indeed the egg? (i.e. was SCO's code even written first?) Who's to say this case should be about a GPL violation?
To imply that IBM's developers are so stupid as to copy and paste code simply begs the question, from their own logic: Who's to say that SCO's developers who weren't so bright and pulled the copy-paste job from Linux? Who's to say that SCO didn't put the source code in there themselves, intentionally. I mean, that scenario seems more likely to me: Linux source code is freely available. Nobody outside of SCO will have seen SCO source - and if they have, they're tied and raped with NDAs.
Anyway, that's enough. Here's to SCO choking on its own arrogance.
There is a possibility that one or more Linux developers got some code from UNIX system V. I think so because not all the programmers are familiar with the copyright law. You see, there is a general belief the patents expire, and that is true, but the copyrights will NOT expire in our lifetime. It is theoretically possible that one particular developer saw the copyright notice on a 25 years old piece of UNIX code so he might think that copyright expired. I do agree there is just a small chance, but it is possible.
Many people ask why SCO do not reveals the code in question. That is because they do not want that code to be removed at this time. Some of us said they will have to show the code in the court anyway and after that the code would be replaced very fast. That is a possible scenario... but what if it will be to late ? I will explain...
I believe the most serious claim is that Linux is an illegal derivative of UNIX. It looks strange to us but I think there is a chance to convince a judge/jury. If some of THEIR code was copied in the key parts of the kernel, let say I/O or TCP stack, and they will show the kernel do not work without that code they have a case... saying that Linux is an illegal derivative of UNIX.
Yes, it is true they may have only few hundreds of lines of UNIX system V in the Linux kernel and the kernel has more than 3 million of lines, but if they show they are in the key parts of the kernel they may convince a judge/jury they own the copyrights for central parts of the kernel and Linux exist just because of their code.
Secondly, IT DOES NOT really matter who owns the UNIX System V copyrights. Even if the owner is our friend now it might turn against us in the future.
We can not simply assume this is going to end like AT&T vs. BSDI because we can not take the risk... they have nothing to lose, but we have.
I think we must act now... I think we have to start auditing the code and to try to find out ANY proprietary code in the kernel before the trial. I do know some may certainly believe that is irrelevant because we do not have their source tree commit dates and the code may be ours anyway. That is certainly right, but I still think we have to try.. at least we will see how many similarities we have and since we know who wrote almost any part of our code we probably clarify most of it.
We will not be able to trace the code which is "trade secret" (developed by IBM and SCO) so some of us will think this is useless. Maybe.. maybe not... because it is not the same think to have 20 lines or 200 lines of proprietary code in the kernel.
I DO NOT want to say they have a case but we have to put the worst things first...
So, who wants to help please write to spaniard@softhome.net to discuss modus operandi.
The RSA patents and Unisys's LZW patents both were valid from my point of view. I don't know if there were patents on Diffie-Hellman key exchange (if yes, they have expired), but that also would have been a worthy software patent. Yes, Unisys mishandled the patent badly, but the original algorithm definitely was non-obvious and innovative.
Borderline cases (that, as far as I know have not been patented): Splay trees, A* algorithm.
Of course nowadays you can get a patent on "Doing X with a computer" and "Doing X with the Internet", where X can be anything from selling candy to taking a dump.
As far as I can tell, while there are some worthy software patents, the vast majority is crap. And even the few valid patents cover algorithms that would have been developed either way, so the patent system is not "promoting science and the useful arts" in the softeware field.
Stephan
No. It's not. You're acting like the guy in "1984" who just couldn't wait to cut words out of the English language. We have perfectly good words to describe various forms of intellectual dishonesty or unauthorized duplication of works without conflating something like copying an mp3 with grand theft auto.
We don't need to resort to pretending that copying something is the same as stealing something for copying something to be wrong.
I do not have a signature