Ransom Love's Linuxworld 2000 Keynote Speech:
Caldera To "Add Components" to Linux Kernel To Make It Scale
SCO's UNIX code may have legitimately reached the Linux kernel, even by means of Caldera itself donating the code. Ransom Love, then CEO of Caldera, gave the keynote speech at Linuxworld back in August of 2000. In the speech, he said that Caldera would itself be donating code to the Linux kernel in order to make it scale for high-end business use, as part of the IA-64 Linux Project, which Caldera was a member of, along with HP, SuSE, Intel, RedHat, SGI, Debian, Dell, NEC, LinuxCare, Cern, VA Linux, Turbo Linux, and IBM.
According to this TWikIWeThey page on the Trillian Project, the original name for the IA-64 Linux Project, SCO/Caldera employees in fact did contribute to the Linux kernel. So did HP, which you can confirm by reading the transcribed Q and A from the press conference held in 1999 when the IA-64 source code was made publicly available.
In fact, there was a Caldera IA-64 Linux distribution, under the GPL, initial release date 8/4/00, listed on the IA-64 Linux Project home page, as "Caldera Systems (initial release 8/4/00) Download at ftp.caldera.com/pub/OpenLinux64", although the link is now broken, If you try to click on the link to "Caldera: Itanium NetFarm available at http://www.caldera.com/partners/developer/ia64.html", on that same page, you get a "Document Not Found" message on SCO's site for that as well.
Obviously, SCO knows about the IA-64 Linux Project. Perhaps they took the page and the ftp down when they decided to stop distributing Linux, or maybe it's just more "disappearing" material from their site.
Whatever the motive, the inevitable question must be: How could SCO say in their Complaint that Linux was just a hobbyist plaything that had no hope of scaling without IBM's support, when they themselves were a leading member of this project that deliberately worked to make it scale and in fact succeeded in doing so? It's very puzzling, unless you are a cynic.
SCO, today's SCO, says their code reached Linux in an unauthorized way, and they blame IBM. But Love in the August 2000 speech says his company, along with IBM, was committed to Linux and specifically committed to making it scale to 64-bit and they were working together on Project Monterey and IA-64 Linux at the same time. He says Caldera would be donating code to the kernel and working with Linus Torvalds to make the kernel as robust and scalable as the Project Monterey kernel and the Unixware kernel.
And the .pdf transcript of the Q and A from the press conference says HP contributed the initial kernel and IBM provided only "performance tools, measurement, and analysis". What other features would the group provide the Linux kernel? "SMP, clustering, large memory, large file systems, and performance monitoring." And this all had Caldera's blessing; they were active participants. And now they are shocked, shocked to find UNIX code in the Linux kernel?
Some of these features are the very features SCO has alleged were made available to Linux by IBM from AIX code, but this IA-64 Project material makes clear that Linux had those listed features already back in 1999, because the project donated their work to the standard Linux tree.
We haven't yet been told exactly what code SCO is talking about, and we have to stick to the facts, but this raises a great many... well, cynical questions in a person's mind. I'm feeling a bit shocked myself. If I were working on this case, I'd surely be making a list of witnesses for the supervising attorney. Now I understand what IBM means in its affirmative defenses, when it lists laches and waiver. I'll be explaining all the affirmative defenses as I find the time, bit by bit.
I have an mp3 of the pertinent section of Ransom Love's speech, a couple of minutes' worth, and here's a transcript I made of it. If you want to hear it yourself, go here and scroll to the bottom of the page:
"Q: What happens about Project Monterey, because that conflicts with the IA-64 Linux, 64-bit Linux?
"Love: OK. I don't -- if we do our job right in making Linux scale over like UnixWare to the degree that everybody, that we know we can... May I ask, some people have said, "Well, people have tried this in the past, but they haven't been that successful," may I suggest: we don't have any ulterior motives for not making it successful. Technologically has not been the reason why it hasn't done it before. There's always some other motive, right? And so to talk about Monterey, clearly we want to make sure we have the same level of Linux integration on Monterey that we would have in our Unixware product. Now, we don't control, I mean, we have a great relationship... it's a joint development relationship with IBM which we intend to preserve ... but they have similar interests and so this is really a very synergistic, uh, this transaction is great for all of the major partners as they have already wanted to embrace Linux moving forward.
"Now, let me address one other aspect of your question, which is that the Monterey Project is in conflict with the IA-64 Linux Project. I don't believe it's in conflict at all. Now, clearly, we have tremendous vested interest in the IA-64 Linux Project and with the acquisition of SCO, they've been doing a lot, so you combine those, and we've got one of the more comprehensive offerings, I believe, on the IA-64 Linux. So that's clearly an area that we're very committed to. But like Unixware, there's elements of the Monterey kernel that are more scalable, OK? Now, on the IA-64 platform, I don't know how long of window that is, but today, it's a little bit more robust and more scalable than the IA-64 Linux is today. Now, I'm not saying that over time that won't change.
"But, and let me address one other thing. Sorry, (laughs) you're getting all of it through one question. But clearly we are going to add components back to the Linux kernel on both IA-32 and IA-64 platforms. We'll work with Linus and everyone in order to make that available. That will take some time. And as I mentioned earlier, I don't know that over time you can have a single kernel -- in fact I know you can't -- that will scale, you know, the breadth of IT technology needs. So I think we're looking, in the Linux community, at having multiple kernels, so...
"Q: Multiple Linux kernels? Or multiple UNIX kernels?
"Love: Multiple Linux kernels as well, over time.
"Q: Thank you.
"Love: You bet.
If you want to hear it for yourself, go here and there is a video of the entire speech. The speech lasts for 50 minutes, and the transcribed portion is about 45 minutes in, about 5 minutes before the end of the speech.
For more on the Trillian Project, now the IA-64 Project, here's where they can be found. And here's a Caldera press release from 1999 about them joining the Trillian Project.
And if you are a little vague on what all this 64-bit stuff is talking about, the TWikIWeThey page and these two Ars Technica articles here and here can help. For historian Peter Salus, author of "A Quarter Century of UNIX", on a history of UNIX and the UNIX licenses, there is video and an mp3.