Help me grok the root of SCO-IBM case (no pun intended)

[Fatt]

I'm not a linux or unix expert and I can barely program in VB, let alone C, so please forgive me if I inadvertently say something silly.


I'm trying to understand the SCO-IBM case.

From what I gather, SCO is complaining about a shared library.

If I understand correctly, when you write a program you include some libraries with the program, and that these libraries contain "tools" that the program needs in order for it to do the things that you want it to do.

Now, I've read SCO's complaint and leaving aside the fundamentally silly notion that IBM must have stolen code from SCO because it would have been too hard for IBM to figure things out by themselves, SCO is complaining about some sort of library that they apparently feel belongs to THEM.


So... can anybody explain to me what this library they are talking about is, and what it does?

And while you're at it, if you see any misconceptions in my question, it'd be great of you could clear them up.

 

[Barnaby W. Füi]

SCO is actually claiming that their code is in the kernel itself.

 

[Fatt]

I read the complaint and it sure sounded like they were talking about a library.

 

[Fatt]

Taken directly from SCO's Summons & Complaint:


The SCO OpenServer Libraries

36.         In creating the thousands of SCO OpenServer Applications, each designed for a specialized function in a vertical industry, software developers wrote software code specifically for the SCO OpenServer shared libraries (hereinafter the ?SCO OpenServer Shared Libraries?).

37.         A ?shared library? is a common set of computer code inside an operating system that performs a routine function for all the applications (software programs) designed to run on that particular operating system.  Thus, Microsoft Windows has its own set of shared libraries.  SCO OpenServer (UNIX designed for Intel chips) has its set of own shared libraries.  Sun Solaris (UNIX designed for SPARC chips) has its own set of shared libraries. 

38.         The shared libraries of all operating systems are designed with ?hooks.?  These ?hooks? are computer code that trigger the operation of certain routine functions.  A software developer can shorten the development effort for any new software program and create a more efficient code base by writing programs that access the various ?hooks? of the operating system, and thereby use a shared set of code built into the operating system to perform the repetitive, common functions that are involved in every program.

39.         Every one of the specialized applications (software programs) designed by various third-party software developers for use on the SCO OpenServer operating system was written to access the various ?hooks? built into SCO OpenServer; and therefore designed to access the SCO OpenServer Shared Libraries. 

40.         The SCO OpenServer Shared Libraries are the proprietary and confidential property of SCO.  SCO OpenServer has been licensed to numerous customers subject to restrictions on use that prohibit unauthorized use of any of its software code, including without limitation, the SCO OpenServer Shared Libraries.

41.         Shared libraries are by their nature unique creations based on various decisions to write code in certain ways, which are in great part random decisions of the software developers who create the shared library code base.  There is no established way to create a specific shared library and the random choices in the location and access calls for ?hooks? that are part of the creation of any shared library.  Therefore, the mathematical probability of a customer being able to recreate the SCO OpenServer Shared Libraries without unauthorized access to or use of the source code of the SCO OpenServer Shared Libraries is nil.

 

[Barnaby W. Füi]

As far as I know, their claims have been against linux only, and not gnu or any other libraries or utilities/apps/whatever. The term "library" is a little vague. The kernel only uses its own source, and does not link against other libraries (afaik), plus, it has to be compiled statically since you can't really link to libraries before filesystems are mounted, etc

 

[Drift3r]

There claim is that they pretend to have a claim but really in all truthfulness they don't have crap.

 

[Nothinman]

They claim people at IBM put code that SCO owns into the Linux kernel without SCO's consent, probably stuff related to SMP and NUMA.

Now, I've read SCO's complaint and leaving aside the fundamentally silly notion that IBM must have stolen code from SCO because it would have been too hard for IBM to figure things out by themselves, SCO is complaining about some sort of library that they apparently feel belongs to THEM.

Not because it would be too hard for them to figure out, but just that they took the code from something SCO licensed to them, probably parts of AIX that came with the orignal UNIX licenses. IBM has been doing a lot of kernel work lately (would be cool to see a break down of where messages on lkml come from, us.ibm.com, dell.com, etc) that has made huge jumps in Linux SMP, NUMA, large block device (>2 TB), large memory (64G), etc. Thing is most of that stuff probably would have moved very slowly if IBM wasn't helping just because it's hard to get ahold of hardware like that, you know how much a several node NUMA-Q cluster with 64G memory costs?

 

[drag]

IBM stated several times that they had no desire to use Unix IP in any open source stuff. They had a different set of developers for both projects and kept them seperate with strict rules on what sort of information goes were.

At least that's what they say.

Anyways SCO is obviously full of it. They refuse to say what parts of the kernel the stolen code is in and the refuse to even hint at what the code could be.

And even if their is similarities it's just to easy to explain away to the fact that their is only so many different ways of solving the same problems and different developers can easily find similar solutions to similar problems. Plus the Unix code goes back a LONG ways. The crap that SCO owns from AT&T is completely obsolete and useless for anything modern. It's just been superseded to many times. Plus their has been much movement of code from company to company including stolen BSD code that was added to commercial unix variants (with no credit towards BSD), plus numerious things like the TCP/IP stack that was added thru legit means....


Anyways there are also gleeming inconsistances like the kernel's SMP support. Linux has OS's that supported SMP BEFORE SCO did. Getting SMP support on 2 proccessors from a single proccessor only systems is much much harder then expanding 2 proccessor support to 8 way or even 32 way systems. The code is already there it just needs to be expanded and have some of the bugs worked out.

Linux has had multiproccessor support since kernel version 2.0.
And guess what company helped that out a lot? CALDERA. Caldera provided the initial code that made smp support possible.....

And that was a LONG time before Caldera ever bought SCO out..... And then changed their name from Caldera to Santa Cruz Operations for the name recognition.

All that is protected under the GPL and SCO's claim has so many holes in it it's not even funny.

All this time of course the stock of SCO is going up. SCO needed the publicity and stock holders are so ignorant that they think that SCO can seize the Linux market (which is now much larger then the Unixware market) thru legal meanderings. Meanwhile the top execs are dumping company stock as quick as they can without being TOO obvious and when IBM/Redhat crushes them they will probably have enough money saved up to retire early and leave the poor investors holding the bag.

 

[ProviaFan]

Drag covered it nicely. Still, I don't see how they can clain Linux "stole" NUMA from them, when AFAIK their UnixWare OS doesn't even have it.

 

[drag]

also another thing I think about...
Would it be easier for a programmer to steal code from a completely alien kernal, make it work well with few problems in a completely different OS, while obscuring it thru creative manipulation, and keeping it a complete secret from the rest of the community who normally do things in the complete open as a team effort. OR simply help out the hundreds of other programmers that are interested in this feature to create it without the intellectual dishonesty?

 

[bsobel]

Anyways SCO is obviously full of it. They refuse to say what parts of the kernel the stolen code is in and the refuse to even hint at what the code could be.

Actually, SCO claims to have given access to the offending code to 80 firms under an NDA agreement. It's actually not 'obvious' that SCO is full of it. What I think is being blown (completely) out of proportion is the amount of damages that SCO claims this infringment has caused/will cause.

Bill

 

[drag]

Now don't get me wrong.. this is just my understanding of the situation, correct me if you see fit:

I don't care about a NDA agreement. It's a trick. How does that give it credence?

Like I said before, if you compare the source code of different kernels designed to do the same thing on intel hardware you will find many similarties. What SCO claims is that people stole code, changed it enough to disguise it, then implemented it on a different OS design.

The NDA agreements are just so that SCO can test it's case out on semi-knowledgeble people. It doesn't give any credence to their case one way or the other. It's just a lawyers trick try to make people think that they may actually have something. For all you know they could be sending the companies copies of MS-DOS source code and nobody would know the difference, exept the people they gave it too, which wouldn't be able to tell anybody anyways.

IBM had has no desire to share Unix IP into Linux. They wanted to get linux to work on their hardware effectively so that they can compete in the low to mid range Unix market that Linux will begin to dominate shortly, if it hasn't already. Everybody else doing it, Sun, Novell, HP , etc etc. Why would SCO pick on IBM alone? IBM has more money and R&D poured into AIX then SCO does. They stand to loose more from "devalueing" unix.

All SCO has to do is post several examples of what they are talking about. This will discredit Linux, force IBM to settle and developers would quickly replace the offending bits and SCO wouldn't have ever have to worry about people stealing their IP.

It's plain as the nose on my face that SCO has some other agenda then just protecting it's source code. It's double level, Linux low-end, Unix high-end marketing sceme didn't pan out and SCO was close to death. Linux is better suited for what SCO traditionally has been used for, and what Unixware can do better then Linux(which since 2.4 has been severly limited) other versions of Unix can do better then Unixware, and at a cheaper price.

Heres another example of Unixware vs Linux capabilities. Originally unixware had the edge in file systems. Unixware could support 1 terabyte max filing systems, however Linux could support up to 2 terabytes. However max file size for Linux was a less impressive 2GB vs Unixware 1 terabyte max file size. So for companies that needed to provide large databases, Unixware was definatly a better solution.

However that was back in 1999, what are Linux limits now? What are Unixware's? I doubt that unixware changed much, but Linux's cababilities have increased significantly. With kernel LVM support on the xfs filing system on 64 bit machines Linux is limited to a paltry 9EB size filing system and a insignificant maximum file size of 32TB.

Linux has more developers, and a large number of realy realy good developers, more market share, and a gigantic userbase to help troubleshoot problems compared to Unixware. Old school Unix is outclassed. If Linux doesn't have what your company needs your better off sticking with Solaris or AIX then settling for SCO. Now that I think about it, it is probably more likely that SCO stole code from linux for many of it's features. I figure that with 2.4 kernels SCO still had a chance to prove it's point that it still held value for higher-end customers, but now that they saw that 2.5 was nearing completetion and 2.6 was coming the finally realised that their version of the 2 tier unix solutions wasn't going to pan out, and they came up with this sceme in a last ditch attempt to make themselves stay relevant.

 

[Barnaby W. Füi]

Originally posted by: drag
What SCO claims is that people stole code, changed it enough to disguise it, then implemented it on a different OS design.

From what I've heard, they claim that there is actually line-for-line copies of their code in linux.

IBM had has no desire to share Unix IP into Linux.

What is your proof of this? If this is just your assumption or opinion, you should make it more apparent, instead of stating it as a fact.

(The above sentence applies to the majority of your post )

 

[bsobel]

What SCO claims is that people stole code, changed it enough to disguise it, then implemented it on a different OS design.

Drag, I'm not going to defend SCO, as I am not under the NDA so I haven't compared the code myself. That said, SCO is claiming the code has been copied line for line (no 'changed enough to disguise it' involved). As for the NDA's, my point was simply to counter your factually incorrect statement of "They refuse to say what parts of the kernel the stolen code". You can say "They refuse to publically say what parts of the kernel the stolen code".

Cheers,
Bill

 

[drag]

hehe, Is their a difference between only publicly admit or not admit all at this stage? Court dates are a long way off.

If SCO REALY REALY wanted to protect their IP all they have to do is simply show Linus and friends what parts have been stolen. A month later they would be replaced and everyone who cares would have a SCO free Linux kernel.. Don't forget that even if they did have line by line code in the Linux kernel, Then why they themselves have been distributing Linux code under the GPL? It's just a little oversight don't you thing?

And anyways about the seperation of AIX and Linux in IBM-land, is not just my opinion. I have read it in several places and Greg Lehey -a freebsd guy- would know a lot more about it then me.

From here

But is the lawsuit justified?
That's one question that's easy to answer: No, in no way whatsoever. If it were, they should at least have made specific accusations as to what IBM is supposed to have done. Having worked on Linux for IBM, I can state categorically that the separation between AIX and Linux is complete. Nearly all of the people working on the Linux kernel have no access to AIX source code. It's theoretically possible that some people do have such access, though I know of nobody, but IBM has guidelines for that case, just to be on the safe side: don't read AIX code and write Linux code in the same place. Read the code, go elsewhere and write. Even this, though, would hardly be useful: AIX is UNIX, Linux is Linux. The kernels have such completely different structures that any code import would be a waste of time: it's easier to write it from scratch.

SCO's talk about SMP scalability is nonsense. They don't have any useful technology in that area. Linux scaled better than UNIX System V long before IBM came on the scene. It's true that IBM is doing a lot of work in that area, but it's based on the PowerPC architecture, and SCO's version of System V doesn't support that platform. So what use would the code be, even if it were a drop-in replacement on the i386 platform?

Here's another thing from Lehey. from a interveiw from Byte magazine

Also this is probably old hat and most of you allready read it, but Here's OSI's position on SCO vs Linux.
I got much of my assumptions from that.

I am not a programmer, but it makes sense to me that it would be easier to right new code from scratch then incorporate it line by line into a completely alien OS? Linux may act like Unix to the end user, but that kernel I bet is run completely different from AT&T decendents... How can Linux steal features and stuff that Linux had BEFORE SCO and have those capabilities in a way that exceeds SCO's? Wouldn't stealing code then making a hack to fit into linux create a inferior version, instead of a superior one?

And your right, SCO did say some line-by-line code was in linux. But they also did charge that lots of stolen IP was obscured by changing it in sneaky ways.

 

[drag]

Also bsodel, I appreciate your responses. I like discussing stuff and I like it when people have good aurguments, even if I end up wrong in the end.

 

[ProviaFan]

IIRC (and this may be wrong; I s/can't/am to tired to/ find a link right now), someone who actually saw the code said it was just a short section, with no line numbers, file names, or enough comments to give anyone a clue where it came from. For all I know (and given how smart the people at SCO seem to be), I wouldn't be surprised to see something like this, over and over again:

int main()
{
/* different code that doesn't matter clipped here */

return 0;
}

"Hey look, they keep copying our code!!! Can't you see?!? That main stuff with those weird curly braces keeps appearing all over the place!" :Q