MozillaQuest Magazine: "83. As long as the Linux development process remained uncoordinated and random, it posed little or no threat to SCO, or to other UNIX vendors, for at least two major reasons: (a) Linux quality was inadequate since it was not developed and tested in coordination for enterprise use and (b) enterprise customer acceptance was non-existent because Linux was viewed by enterprise customers as a "fringe" software product." [Caldera v IBM Complaint]
MozillaQuest Magazine: Is any of this true?
Alan Cox: Linux for enterprise definitely predates IBM involvement. The testing, the development and the focus on Linux server systems was occurring in Linux companies, with folks like Dell and their customer base providing incentive and financial reasons to do it.
Richard Gooch: Addressing (a): Linux was being tested before IBM got interested. Addressing (b): I think IBM got interested in Linux because it was already coming out of the "fringes".
David Weinehall: Sure it may have been that Linux was no threat to SCO while uncoordinated and random, but the coordination was not done by IBM; it has evolved through time by the people on the lkml. [linux-kernel mailing list] We're talking the 0.9 days here, probably.
MozillaQuest Magazine: Is Linux kernel and GNU/Linux operating system development "uncoordinated and random"?
Alan Cox: Yes and no. The power of Linux is that it is self co-ordinating and that it explores random fringes as well as the more direct paths. We still develop enterprise stuff the same way. IBM has had code refused over stuff from random users with crazy ideas that prove the right solutions.
What has to be co-ordinated is *QA*. Software QA is a well understood, non Unix-centric, and alas an inadequately-practiced art.
Richard Gooch: Yes and no. It's complicated, and works in ways that aren't familiar to traditional corporate models of software development. It's not a question that can be answered in the traditional corporate framework.
David Weinehall: Not really, no.
MozillaQuest Magazine: Was "Linux" a '"fringe" software product."' before IBM rescued GNU/Linux and the Linux community?
Alan Cox: IBM made a difference to its perception, but that's not from IBM code. IBM saying "This Linux thing is cool" is what made that difference, (along with Oracle and a lot of other big names saying that too.)
Richard Gooch: Ha! As I said above, it was already coming out of the fringes. And Linux wasn't in need of rescue. It was doing quite nicely. IBM saw a new market trend and decided it wanted a piece of the action. IBM, to its credit, is adapting rather than hanging on to a shrinking market.
David Weinehall: I don't know if I even should gratify that one with an answer :-)
MozillaQuest Magazine: "84. Prior to IBM's involvement, Linux was the software equivalent of a bicycle. UNIX was the software equivalent of a luxury car. To make Linux of necessary quality for use by enterprise customers, it must be re-designed so that Linux also becomes the software equivalent of a luxury car. This re-design is not technologically feasible or even possible at the enterprise level without (1) a high degree of design coordination, (2) access to expensive and sophisticated design and testing equipment; (3) access to UNIX code, methods and concepts; (4) UNIX architectural experience; and (5) a very significant financial investment." [Caldera v IBM Complaint]
MozillaQuest Magazine: Was "Linux" "the software equivalent of a bicycle" while "UNIX was the software equivalent of a luxury car" before IBM rescued GNU/Linux and the Linux community?
Alan Cox: Linux was more the earlier Japanese cars. Not so comfortable, didn't come with cup holders but cost a lot less and unlike proprietary Unix (the US made car of the time) you didn't have to look in the rear view mirror for pieces falling off all the time.
Richard Gooch: No. It's a stupid analogy, and I hope that in court, it is seen as such. But, if pressed to continue with the analogy, Linux was more like a high-performance racing car. What the Linux community is doing now (and is a topic of ongoing, often heated, debate) is keeping the powerful engine, robust suspension and responsive steering which make it so great, and is adding lightweight padded seats and compartments for optional luxury extras such as Hi-Fi sound and heads-up-display GPS navigation for those who like the feel of a luxury car but want the power of a race-car.
There is one point that SCO raises which is correct in some sense: (5). There has indeed been a significant investment in Linux. However, it wasn't IBM that made the investment. It was the Linux community that made the investment, much of which was not remunerated, as it was volunteer work.
David Weinehall: Obviously not.
Of course, most people can fix their bike on their own, because a bike is simple, and this goes for Linux too; not because it's simple but because it's free and all source available. And Luxury cars cost a lot, so does SCO Unix. But that's about all I find significant to this likening.
I'd attribute most of the heavy-weight redesign of the kernel to Alexander Viro, Theodore T'so, David S Miller, Linus himself and a few select more; none of them IBM employees afaik [as far as I know].
MozillaQuest Magazine: Did the Linux kernel and GNU/Linux developers and groups lack the technological capability of producing an enterprise level Linux without being bailed-out by IBM?
Alan Cox: If you look at the Linux engineers you'll find Ph.D. level researchers, experts in their fields, and senior members of standards committees.
Richard Gooch: We weren't "bailed out". Linux was developing nicely anyway. IBM's interest in Linux is appreciated, and showed considerable vision, but Linux would have continued developing regardless.
David Weinehall: Obviously not. Most/all technology in the Linux kernel is based on well-known operating system theory, published in USENIX papers, taught in OS theory etc.
MozillaQuest Magazine: "85. For example, Linux is currently capable of coordinating the simultaneous performance of 4 computer processors. UNIX, on the other hand, commonly links 16 processors and can successfully link up to 32 processors for simultaneous operation. This difference in memory management performance is very significant to enterprise customers who need extremely high computing capabilities for complex tasks. The ability to accomplish this task successfully has taken AT&T, Novell and SCO at least 20 years, with access to expensive equipment for design and testing, well-trained UNIX engineers and a wealth of experience in UNIX methods and concepts." [Caldera v IBM Complaint]
MozillaQuest Magazine: Is this true? I thought the Linux kernel and GNU/Linux can handle 32 CPUs?
Alan Cox: Linux 2.4 is happy up to about 4 CPUs, true. Ingo Molnar did the scheduling work to fix that. People like Jens Axboe did the core work on fixing the block I/O layers to deal with it. That's Red Hat and SuSE. People like Larry McVoy (an ex-SGI guru) provided a lot of ideas that led to some key optimizations too. Very large memory systems have been very much an IBM thing, along with decent NUMA support.
Work != [does not equal] Work well. I can construct workloads where 32-CPU Linux scales beautifully with the 2.4 kernel. Unfortunately if you try a real workload on 2.4 with that many CPUs (say oracle) the numbers look less ideal.
Richard Gooch: Maybe it took AT&T, Novell and SCO 20 years to figure it out, but the Linux community took just a few years, without access to proprietary IP. As I said earlier, Linux was running on at least 14-way SMP systems before IBM joined in.
David Weinehall: Yes. And SGI has likely provided more help with this than IBM has.
Make that 32 for the v2.5 kernel.
This isn't even about memory management, but about scheduling. Says everything, doesn't it?
20 years ago, there weren't even any 32-processor computers.
MozillaQuest Magazine: "86. It is not possible for Linux to rapidly reach UNIX performance standards for complete enterprise functionality without the misappropriation of UNIX code, methods or concepts to achieve such performance, and coordination by a larger developer, such as IBM." [Caldera v IBM Complaint]
Alan Cox: IBM has been delivering that kind of technology to people since before SCO existed, before Unix existed.
Richard Gooch: Well, sorry, SCO, but we figured it out on our own.
David Weinehall: Pah.
MozillaQuest Magazine: Did the Linux kernel and the GNU/Linux operating system developers and groups lack the technological capability to "reach UNIX performance standards" and of producing an enterprise level Linux without being bailed-out by IBM?
Alan Cox: IBM helped, no doubt about it. It might have taken a little longer without them, but it's all in books and papers.
Richard Gooch: No.
MozillaQuest Magazine: SCO-Caldera claims that the Linux kernel and/or GNU/Linux source code is based upon its UNIX source code. In paragraph "74" of its Complaint Caldera d/b/a SCO alleges: "A new operating system derived from and based on UNIX recently has become popular among computer enthusiasts for use on personal, educational-based, and not-for-profit projects and initiatives. This operating system is named Linux."
Richard Gooch: False. Linux contains no SCO source code. Frankly, even if offered freely, we probably wouldn't want it, as it's likely to be seen as bloatware.
David Weinehall: No, inspired by UNIX and conforming to Posix.
MozillaQuest Magazine: Prominent members of the Linux kernel, GNU/Linux, and UNIX communities have denied SCO-Caldera's claims that Linux was derived from UNIX in our SCO IP Part 2 article -- Linux was built from the ground up and independently of the UNIX source code. Do you have anything you would like to add to that?
Alan Cox: Derived from - no. Inspired by - certainly.
Richard Gooch: Not really. It's a pity the weasels at SCO are litigating instead of working for a living (such as making products, which is what tech companies are supposed to do).
David Weinehall: Well, if the AT&T source code had been free, Linus would probably not have developed Linux at all...
MozillaQuest Magazine: How about the implication that the Linux kernel and the GNU/Linux operating systems are more like toys than a serious OS?
Alan Cox: Once upon a time.
Richard Gooch: It was a toy once. That stopped being true many years ago. It was no longer a toy before the time IBM got involved. It's been a real OS for many years. I think millions of users agree with me.
David Weinehall: Compared to, say, NT? Hahahahaha...
MozillaQuest Magazine: Do you have any other thoughts or comments that you would like to add to anything regarding the Caldera v IBM lawsuit and/or the SCO-Caldera IP issues?
Alan Cox: Not really. My guess is that SCO went to IBM and said "buy us" and IBM said "no". There may be contractual issues between SCO and IBM over the failed Monterey project between the two Unix vendors. I've dealt with IBM lawyers and given their paranoia, I suspect IBM was very careful about such things. That bit isn't really a Linux thing . . .
Its kind of sad to see that proprietary Unix vendors who killed Unix by fighting each other, plan to fight to the death and go out without learning one single lesson.
Richard Gooch: What SCO is doing is unethical, and they know it. They will generate a backlash of resentment. If they had simply sought to generate revenue from licensing their UNIX libraries, no-one would be taking issue, and many would have wished them luck in trying to adapt to a changing market. They've shot themselves in the foot. Stupid, stupid, stupid.
Please see the first two parts of our series about SCO-Caldera's IP claims plus its intentions to enforce and license its intellectual property rights.
Related MozillaQuest Articles
SCO-Caldera v IBM: