SCO has responded [PDF] to Novell's Notice just filed in Utah District Court. Novell is asking that court to assign both the Novell case and the IBM case to the same judge or to consolidate the two cases. SCO's Chapter 11 Trustee, Edward Cahn, also submits a statement in support of SCO's position, which is basically that it needs to hurry up and finish the Novell trial. It's running out of money. It doesn't want to detour to decide the SUSE arbitration first, which Novell also is requesting, and it sees consolidation with IBM to be not only not necessary but something that will drag things out.
Granting the Novell request for the same judge to be assigned to both cases surely wouldn't slow down the Novell case at all, but it would help the new judge having to suddenly deal with these complicated and interwoven cases to at least comprehend how they interact, particularly with respect to the GPL.
Don't forget that Judge Dale Kimball was assigned to both IBM and Novell. So assigning the two cases to one new judge to replace him would simply be returning to a kind of status quo. And if Novell prevails in the arbitration, SCO's copyright claims pretty much go poof. Why wouldn't you want to do that first, particularly since now it's ripe on the vine, whereas when the Utah court decided to go forward on both tracks, the arbitration was just beginning. It would have been decided long ago, had it not been for SCO's bankruptcy filing.
Neither SCO nor Mr. Cahn show any comprehension, from what I could see, of the General Public License and how it will affect SCO.
Here are the filings:
Do they not understand the GPL, I wonder? The most important fact is that SCO knowingly distributed UnitedLinux after it sued IBM, a distribution which included all the code at issue in the IBM litigation, which it distributed under the GPL, which is a copyright license that allows anyone in the world who receives it to copy, modify and distribute the code in harmony with the license terms. Code SCO is suing IBM over was in UnitedLinux, such as JFS, and SCO knew it at the time of distribution, because it mentioned it or other "high end enterprise features" in ads, white papers [PDF], press releases, and presentations.
11/24/2009 - 602 - RESPONSE re 600 Notice (Other), filed by Plaintiff SCO Group. (Hatch, Brent) (Entered: 11/24/2009)
11/24/2009 - 603 AFFIDAVIT of Brent O. Hatch in Support re 602 Response (NOT to motion) filed by Plaintiff SCO Group. (Attachments: # 1 Exhibit A, # 2 Exhibit B, # 3 Exhibit C, # 4 Exhibit D, # 5 Exhibit E, # 6 Exhibit F, # 7 Exhibit G)(Hatch, Brent) (Entered: 11/24/2009)
And may I point out that SCO is not the copyright holder of that code? It didn't write any of it or even buy it from someone who did. For example, IBM wrote JFS and donated it to Linux from OS/2, not from AIX. SCO has a theory I trust Cahn didn't get to yet, regarding methods and concepts, but the Utah court already ruled SCO can't go into that at trial, even if it had been from AIX.
Here's a paper written by John Terpstra in
2000, before SCO hired him briefly while he was working for Caldera in 20021, about UnitedLinux, and the technical paper lists some of what was in it back then:
jfsutils-1.0.24: utilities for managing IBM's JFS (journaled file system) under Linux... So they knew. Everyone knew and everyone knows. It's just the courts and the judges trying to grasp what is as obvious as gravity to the tech community.
The kernel is based on linux-2.4.19 with enterprise features enabled. The facilities and capabilities of the kernel include: ...
Memory: NUMA, Memory Extension Technology (MXT), Large Memory Support (64GB physical RAM)
And Terpstra added:
The UnitedLinux product, which is the base of all its partner distributions, is identical across distributions. The startup process, core libraries, default tools and utilities, and so on, are now exactly the same for SuSE, Conectiva, Turbolinux, and SCO Linux products.Catch that about core libraries, tools and utilities? And note that SCO then intended to release UnitedLinux for IBM S/390 and PowerPC? Terpstra adds that it already ran on a "quad-processor Intel Xeon-based IBM system with an IBM ServerRAID storage system." Ironic, but true. Now SCO wants to sue IBM for the same code UnitedLinux was extolling as a plus and making sure would run on IBM architectures.
Because the base product is LSB compliant, OpenI18N compliant, and GB18030 compliant, developers who choose a UnitedLinux powered platform will have a well-defined, uniform development platform and a sure and certain deployment solution for their customers. While at the outset the UnitedLinux initiative is focused on the Intel processor architecture, UnitedLinux will eventually release a fully scalable uniform product for Intel IA32 and IA64, AMD x86-64, PowerPC, and IBM S/390 processors.
After the main body of the paper, he listed development tools for UnitedLinux:
Looking to dive in and get started? Check out these development tools for UnitedLinux:
The paper links to everything on the list, but I want you to look at the first link. What do we find? binutils. If you investigate, you will find that other things SCO is suing about, like ELF, are in binutils, and it is under the GPL. Here's the binutils index. Caldera at one point maintained the ELF registry for the GNU binutils project, so they knew and they know. They just say these things. Here's SCO distributing skunkware in 2006, also here, which included binutils and hence ELF and the header files SCO is suing over, and it distributed it all under the GPL long after it sued IBM. Anyone in the world can use ELF. You can read about that here. Yet SCO had the gall to list binutils in this Exhibit [PDF], claiming to be the copyright owner and stating it never authorizd its release outside of UNIX:
1. SCO, as the copyright owner of source code and/or documentation upon which the following files and lines of code were copied or derived, has never contributed or authorized these lines of code or the documentation related thereto, for use in Linux as specified under part 0, or any other provision, of the GPL. You can see from the above evidence what I think of those claims. SCO's expert, Thomas Cargill, also listed ELF, claiming it had no right to be in Linux. But SCO put it there itself, in its own distribution of Linux, more than once, knowing what it was, and under the GPL. You can probably guess what I think of SCO's experts. What exactly is meritorious about SCO's claims? They are, to me... I started to say laughable, but they're not funny. It's disturbing. Perhaps Mr. Cahn can be more specific. No. Really. What exactly is meritorious about any of this?
2. SCO, as the copyright owner of source code and/or documentation upon which the following files and lines of code were copied or derived, has never granted a license to any party that knowingly authorized use of these files or lines of code outside a UNIX-based distribution.
Here's where you'll find Ralf Flaxa's Declaration in the IBM litigation. He says clear as a bell that while he was at Caldera, he knew that things like ELF were in Linux:
28. I understand that SCO claims that certain materials in Linux infringe SCO's alleged intellectual property, specifically: header files required by the Open Group's Single Unix Specification (SUS), header files relating to the Streams technology, and files and specifications relating to the Executable and Linking Format (ELF). Caldera distributed binutils in OpenLinux too. You'll find a complete list of everything that was in OpenLinux eBuilder Enterprise Edition in 2000 here, also distributed under the GPL. You don't have a choice with binutils. And SCO was still distributing Linux, the kernel, in 2006, three years after suing IBM, and without any notice that it was just for their prior customers. Here's where it was distributing ELF under the GPL from its website in 2006 in binutils. That too was distributed under the GPL by SCO, so everything it is suing over it distributed itself, before, during, and after claiming it included infringing materials. Maybe that's why the tech community just laughs at SCO's allegations. They are, to us, simply ridiculous. Now one must hope that Cahn will find out about all this. SCO may not share that hope.
29. While employed at Caldera, I was aware that this material was present in Linux. I know so because of my familiarity with Linux and also because Caldera incorporated it into its Linux products.
They don't seem to realize yet what that all means. Cahn spoke to all the lawyers in the case, but who he really needs to talk to is the Software Freedom Law Center's Eben Moglen, which represents FSF, the author of the GPL.They could explain all this to them better than I can, but I have explained it anyway, as best I can, and with quotations from Moglen and others specifically in connection with SCO's claims, here. Not to put too fine a point on it, but I believe it trumps all copyright infringement claims against Novell and end users. IBM is an end user. Contracts are a separate matter, of course. But if Cahn is thinking about end users and copyright infringement -- and I don't know that he is or isn't -- that's a fool's errand, and pursuing it will only waste more assets, I would suggest.
Of course, SCO wants to press its advantage. The Appeals Court gave it a gift, wrongly in my view, and SCO's lawyers want to keep that momentum going. And they'd like to hurry up before the Supreme Court can do anything should Novell be moving forward on asking that the appeal court decision be reviewed. But it's obviously prejudicial to Novell and to IBM to rush, because while Kimball wouldn't be unduly impressed with the ruling from the appeals court, knowing the facts of the case as well as he does, a new judge is bound to be. Even Cahn is. How would any of them know any better? These litigations have been going on for half a decade. You can't get up to speed in a few weeks. You just can't. So, I see an attempt to stack the deck a bit here. Of course, one could argue that is what lawyers are supposed to do, and to a point, that's true.
There is a perception among those of us who are not new to the case that the appeals court didn't look deeply enough or take the time to get it right, at best. But you'd have to know all these fine details to see that, I think, and the new judges won't have spent every day for years doing what we've been doing. And the more they hurry the trial, the less time the new judge has to get up to speed. Perfect for SCO, but not for truth. From SCO's perspective, though, it doesn't have a choice to be leisurely now. It's running out of money.
Here's something else that I noticed in Cahn's remarks, attached as an exhibit. He mentions that among the things he looked at were "confidential" materials that Boies Schiller has. That has me wondering if they, as has been their wont, intend to spring something at the last minute again, despite being sanctioned for trying that very trick in the IBM case.
Mr. Cahn needs to get to know the Boies Schiller guys just a little bit better, methinks. He is new to this case, after all.
The way SCO tells the story about what's happened to date in the Novell and IBM litigations and the SUSE arbitration is sort of accurate, but not precisely so, at least not as I remember the facts. For example, SCO tells the court that SCO's claims against IBM include copyright infringement claims arising out of IBM's activities related to Linux. Kind of, vaguely, but it's way more complex than that. SCO did get permission to amend its complaint to add a copyright infringement claim, but not about donating to Linux. SCO argued with fervor that its copyright claim was about continued distribution of AIX after SCO allegedly terminated IBM's license. The court didn't believe that. But that is what SCO then claimed. IBM was pulled in to the group of 1500 Linux end users that got the letter about using Linux being infringing, not writing it. Look at SCO's Second Amended Complaint for yourself. And IBM's recounting of SCO's claims in its Second Amended Counterclaims, which also includes this gem:
28. The Linux kernel is subject to the GPL as it is comprised of programs and other works that contain notices placed by contributing copyright holders permitting distribution under the terms of the GPL. The Linux developers' public agreement to apply GPL terms expresses in a binding legal form the conscious public covenant that defines the open-source community -- a covenant that SCO itself supported as a Linux company for many years. Is SCO suing IBM over copyright infringement of JFS, which IBM itself wrote? Or ELF, which SCO doesn't own but even if it did it distributed under the GPL? IBM has counterclaimed for copyright infringement of its GPL-licensed code in Linux that SCO infringed by continuing distribution after it repudiated the GPL.
29. SCO accepted the terms of the GPL by modifying and distributing Linux products. By distributing Linux products under the GPL, SCO agreed, among other things, not to assert -- indeed, it is prohibited from asserting -- certain proprietary rights over any programs distributed by SCO under the terms of the GPL. SCO also agreed not to restrict further distribution of any programs distributed by SCO under the terms of the GPL.
Also the SCO document describes the SUSE arbitration claim like this: "... the Novell subsidiary claims that SCO surrendered its copyrights as part of the joint venture called 'United Linux'." Again, kind of, but not exactly. Here's some info on the SUSE claims, which include:
46. In particular the United Linux members agreed that each member would have an irrevocable, perpetual, and worldwide license to use and unlimitedly exploit any intellectual property rights of the other members in the UnitedLinux Software, which would be transferred to the LLC for this very purpose.... And you'll find here the substantive relief SUSE is asking for, which you will also find on page 54 of the Declaration of Michael A. Jacobs [PDF] [PDF], in support of Novell, Inc.'s Motion to Stay Claims Raising Issues Subject to Arbitration [PDF]:
1. Declare that Respondent is precluded under the Master Transaction Agreement (MTA) and the UnitedLinux Joint Development Contract (JDC) from asserting any copyright infringement claims related to SUSE Linux;
SCO tells the court about the claims to make them sound ridiculous, in my opinion. Well, since when does SCO tell the whole story? And it tells the court that SCO's "principal prospect" for survival is its litigation "assets". Well, I hate to break it to them, but if that is it, then that is it.
2. Declare, in particular, that the MTA and JDC divest Respondent of ownership of any alleged intellectual property rights in any part of software included in the UnitedLinux Software (other than Pre-Existing Technology and Enhancements);
3. Order Respondent to refrain from alleging publicly or against third parties that the use and distribution of SUSE Linux infringes upon Claimant's copyrights, as precluded by the MTA and JDC;
4. Order Respondent to pay damages in an amount to be determined for breach of the MTA and JDC by improperly asserting claims against Claimant and its licensees, and by attacking and withdrawring support for the UnitedLinux project;
The SCO filing also says that IBM brought over a dozen counterclaims, and then voluntarily decided to drop several of them. Again, it's kind of true, but sounds a little misleading to me, in that IBM dropped the patent claims because it was clear SCO wouldn't have any money at the end of the road, which certainly proved to be the case. Even in this filing, SCO says it is "cash strapped". IBM didn't drop them because they thought they'd lose. They just didn't want to spend money on something that wouldn't bring in money. Say, SCO should give that strategy a try. And also IBM stated that it wanted to hurry up to trial and get it over with. Back then, SCO was trying, we all thought, to avoid a trial and if you recall it was arguing for tons of extra depositions just about the patents and hinting at asking to separate the patents issues from the rest, and IBM wanted to get the show on the road. Here is a snip from the discussion between Judge Kimball and SCO's then-attorney handling the patents claims defense, Frederick Frei, at a hearing on June 8, 2004:
THE COURT: It sounds like you're making an argument to separate the patent claims from each other. Here you have a judge and a skilled attorney agreeing that appeals courts don't always get things right, because they are in the clouds, not in trial courts, where the action is going down. And that is true.
MR. FREI: I am not making that argument now. All I can say is that in many, many cases, that ultimately happens, not only separating patent claims from each other, but within a patent the federal circuit has said that the preferred way of trying these cases is to trifurcate. Validity, damages, infringement, done in separate trials. But we're not getting to that. We're just saying, take us away from the rest of this case.
THE COURT: Appellate courts are often fond of saying things that don't work very well in the actual, real trial world, are they not? Instead of 14 trials, have 20 trials.
MR. FREI: They have good intentions.
THE COURT: Yes.
MR. FREI: But these counterclaims were first filed August 6. The counterclaims were amended late September. They were amended again the end of March when originally there were four patents and it was reduced, IBM dropped three -- or dropped it down to three patents. Discovery is barely beginning on the patent side of the case. There was a stay of discovery in the entire case for three months from early December to early March. We served document requests. IBM responded a month ago. We've not yet gotten document requests. IBM served document requests. Our responses to those requests are due next week. There have been no patent depositions taken yet. And yet, IBM says this whole case can be handled to the completion of fact discovery by August 6 of this year. Patents and the rest of the case. That is unconscionable. It's not possible. We have estimated based on our experience that the trial of a patent claims could take up to five weeks of actual trial time. We have estimated that it could take 60, 70 depositions. Most of the reason for that is --
THE COURT: Why would the patent aspects of this case take five weeks to try?
MR FREI: Because we have -- validity is a separate line of inquiry with all sorts of third parties where we have to dig out the prior art, put those witnesses on the stand either live or through depositions. We have upwards of 40 claims that may be at issue. We don't know how many IBM is ultimately going to choose. We have at least four products, separate and distinct products. And we have raised many defenses, I mean, the key which are unique to patent cases : Validity; noninfringement; doctrine of equivalence; enforceability, due to what we say is inequitable conduct in prosecuting the patents; latches [sic], estoppel; and waiver, to name a few of the defenses. And I would --MO
MR. FREI: Right. But these are separate patents, separate inventors, separate documents, separate witnesses. Everything about these patents is separate. Normally, you might have three separate patent infringement suits, and here we have one, and it's just one of 14 counterclaims that's been asserted.
On and on it goes, SCO arguing about needing more time and discovery, and IBM saw the strategy and nipped it in the bud. Patent litigation tends to go on forever, anyway, because there are always appeals and then appeals of appeals, because patents live down the rabbit hole, where things rarely happen with the logic one would expect, so it's worth appealing everything. There is little logic to patent law in the US, which is just one reason patent litigation is so expensive.
SCO is equally misleading, in my view, in how it describes the Novell case, telling the new judge -- who wasn't there and won't know it's not a complete description -- that Novell originally asked for more than $30 million, but only won $2.5:
After a four-day trial in April 2008, Judge Kimball issued findings of fact and conclusions of law awarding Novell approximately $2.5 million (instead of the more than $30 million Novell had initially sought). What really happened was Novell entered the court on the first day of trial and told the court it wasn't going to pursue some claims, so it was reducing what it was asking for to around $20 million:
MR. JACOBS: Your Honor, in this bench trial, we seek a recovery of exactly $19,979,561 from SCO, based on the Court's earlier findings of breach of fiduciary duty, conversion, unjust enrichment. This is an amount we seek from SCO based on its licenses to Sun, Microsoft and what we're referring to as the SCOsource licenses. On the last issue, Judge Kimball didn't think SCOsource licenses to folks like EV1 were about licensing software code, essentially, but rather about not getting sued. I think he was wrong about that, in that I think it was both, but that is what happened. In fact, I hope that issue comes up at trial, as long as they are going to have one. I hope the issue of incidental use also comes up again, because I think Kimball got that wrong too, not to mention SCO's laughably disingenuous the-tree-is-the-branch-too argument about UnixWare and UNIX being the same thing. And I hope the GPL comes up, and how SCO violated the GPL by demanding Linux end users buy a SCOsource license. Maybe they'll leave that to IBM later. But my point is, SCO never tells the story with all the details right and/or complete, and in that way they make themselves sound better than what really happened.
SCO makes much of Cahn's saying to the bankruptcy court that he thinks SCO's claims are meritorious and should be pursued. Again, that is kinda true, but what he actually said to the court was that he was still evaluating everything but to date that was his view. Just not the same thing. And with all due respect, he seems to place undue weight to the appeals court decision, which I think he'll find over time is not warranted. And the appeals court didn't rule on the merits of the copyright claims, only that it was an issue for a jury. "We take no position on which party ultimately owns the Unix copyrights or which copyrights were required for Santa Cruz to exercise its rights under the agreement,” the court wrote. “Such matters are for the finder of fact on remand.” The court also wrote, "We recognize that Novell has powerful arguments to support its version of the transaction, and that, as the district court suggested, there may be reasons to discount the credibility, relevance, or persuasiveness of the extrinsic evidence that SCO presents." That's if it is even appropriate to look at extrinsic evidence in this fact pattern, which is the issue Novell would like the Supreme Court to rule on.
All of which is to say that Cahn's earliest pronouncements are puzzling to me. What exactly is meritorious, if you read the appellate decision? I've learned to wait and see how things play out, though. Things in the SCO universe are not always what they seem. Perhaps Mr. Cahn is hoping for a deal, and so wishes to present a strong position to negotiate from. If you are hoping for a settlement, you don't walk in and announce, So, I'm holding nothing. What've you got?
Still, by saying what he did, he has changed the atmosphere in one significant way. I no longer see it as possible for SCO or the Boies Schiller firm to be sanctioned for bringing frivolous claims. His statement nukes that possibility, and I'm sure there is great relief about that.
And just a reminder:
Monday *Tuesday* is the day for the status conference before the new judge currently assigned to the SCO v. Novell case in Utah:
01-Dec-2009 SCO vs Novell  
Notice of Hearing: (Notice generated by Chambers/slm) Status Conference
set for 01-Dec-2009 01:30 PM in Room 142 before Judge Ted Stewart
Update: I forgot SysVinit, which SCO is also suing IBM over. As it happens, I own Caldera's OpenLinux 2.3-16, among others, which means they can't sue me ever for using Linux, and of course it also means, I think, that if SCO ever did get the copyrights it seeks, I'd still be within my rights to share it, copy it, modify it, etc., since I got it under the GPL. That is why I don't need a SCOsource license, because I already have a license to use Linux, and in fact if I bought one, it'd be a violation of my license, the GPL. I'd check with my lawyer before I did it, but multiply me by millions who will certainly do the same thing with their distributions, and you see the GPL pickle SCO is in. But leave that aside, what is in there? Things SCO claims it never authorized to be in Linux.
If you take a look at the SRPM folder, you will find some of them. For those who don't use GNU/Linux systems, I'll just explain that RPMs are packages, like OpenOffice on Linux arrives as an rpm on systems based on Red Hat, which is what Caldera did base its Linux distro on. It's kind of like .exe in Windows. SRPMs are the same thing, except they are the same packages in source form, not binary. With that background, here are some things you will find in that source folder:
SysVinit-2.76.3-2.src.rpm There it is, our friend binutils, and SysVinit too, among other goodies. The packaged coas means Caldera Open Administration System. Lots of goodies in there too. What does this mean? It means it knew and it distributed the source of these packages, giving everyone the right to use that source, modify it, copy it, and distributed it. Telling prior customers they couldn't use Linux except in binary form was thus a repudiation of the GPL license, which SCO claims falsely it never violated. By my reading, it did. See why suing over these things is just fanciful? There's more on SysVinit here, where we list all the others versions of its distribution SCO/Caldera put SysVinit into and then distributed under the GPL. For example, in OpenLinux Server 2.3's package list of source packages, you'll notice COAS, binutils, iBCS and extra shared libraries, all the stuff SCO is claiming was put into Linux inappropriately and without authorization. But SCO did so itself. Hard to sue yourself, but Boies Schiller may find a way before this is over. It would be the only appropriate ending to this ridiculous saga.
On SCO's experts, one of the individuals SCO hired in 2005 to an expert in the IBM litigation showed up on Groklaw in 2007, the day Judge Kimball ruled against SCO in August. I asked him why he agreed to work for SCO, and he gave a number of answers that you'd expect, but here are the four reasons that I will always remember:
A career move. As it happens, he was supposed to testify about methods and concepts, but that was ruled out, so he won't be taking the stand, I gather, after all. But can you imagine what IBM would do to him on the stand, with a list like that?
- Chance to learn more about UNIX and Linux....
- Helped me make additional contacts so I can pursue a career as an expert witness.
Update 2: There is a review of Caldera Network Desktop v 1.0, one that ran in the December 1, 1995 Linux Journal, by Roger Stafford. You have to be a member of acm Portal to get it now, but if you do, you will find the following paragraph in the review, regarding ELF:
Although it isn't yet possible to upgrade from one release to another, Caldera says that "such tools are planned for the final 1.0 release." [Those tools are in Preview 2, which has just been released. -- Ed] Something Caldera calls the InfoTrack database support system will become part of the overall technical support. OpenDoc support is in the offing. And ELF work is under way. [Again, Preview 2 is based on ELF -- Ed]
So, 1995. They knew what was in there. They were working on it being in there.
Update 3: Look at this, will you? It's Christoph Hellwig announcing publicly that the Linux-ABI 126.96.36.199 release for Linux 2.4.14 is available. Here's what it supported:
The current release supports the follwoing operating systems:
But notice the email he used: firstname.lastname@example.org -- yes, from Caldera itself came the news. What does that mean? In addition to the obvious that SCO needs to sue itself, it means that if folks had patches to send, they would send them to him at his work address at Caldera.
System V Release 3 (SVR3)
Currently only the Intel/i386 platform is supported.
System V Release 4 (SVR4)
SCO Unix 3.x
SCO Xenix 386
SCO Xenix 286 (with userspace x286emul)
SCO UnixWare 7/Caldera OpenUnix8
Sun Solaris 2
ISC Interactive Unix
Since SCO is suing over the Linux-ABI in Linux 2.4 and beyond, they might want to explain how they can sue people for using something they themselves put into Linux when they were a Linux company, Caldera. And did you notice the date? 2001. And here's where you will find the
kernel maintainers for 2.4.18, and there he is again, maintaining linux-ABI but also the SYSV file system, same email address. And in 2001, here he is offering a patch to the kernel from his work address at Caldera, a System V filesystem update:
Update 4: Since SCO phrases SUSE's position in the arbitration as a claim, I assume they are contesting it. Here's what they wrote:
the appended patch updates the System V filesystem driver to the latest
version from Alan's tree.
The follwoing updates are included:
add SCO fast symlink support (me)
- add readonly SCO AFS support (me)
- be more graceful in the case of a wrong filesystem type (aeb)
D. SuSE Arbitration.
I found a PowerPoint presentation SCO created to sell UnitedLinux. It's called "SCO's Vision for Linux, UNIX & UnitedLinux - A Uniform Linux for Business." Guess what it says? That The UnitedLinux, LLC owned the IP. Here's a screenshot of the slide, slide 28:
With respect to the SuSE arbitration, Novell’s subsidiary claims that SCO surrendered its copyrights as part the joint venture called “United Linux.”
1 I have heard from John Terpstra by email with some corrections:
The article you have linked to ... was written in
2002 during my final days at Caldera International (already calling
itself SCO Group).
Darl McBride and others I had worked with were well aware of this
article and it had been carefully reviewed inside Caldera International
before it was released for publication.
So, it was NOT written in 2000, but rather in 2002. The timing is
significant because (as I recall) United Linux officially launched on
November 15th, 2002. This article could not be released prior to
confirmation that the product would indeed release on the scheduled date
of 11/15/2002. That confirmation came only a day or two before my
departure from Caldera (SCOG). My last day at Caldera (SCOG) was 10/31/2002.
The concept of a unified business Linux distribution started in November
1999 at the Comdex event in Las Vegas. Dirk Hohndel was the first
person outside of TurboLinux (whom I worked with at that time) with whom
I discussed the possibility of getting all the Linux distributions on
board. He was CTO of SuSE at the time. Together we polled the Linux
vendors at Comdex 1999 to commence the initiative.
When I left TurboLinux October 31, 2000, to join Caldera, Ly Pham (then
CTO of TurboLinux) asked me not to drop this important initiative. I
faithfully followed through on her request with the full support of
Ransom Love and the rest of the Caldera International (later SCOG)
My purpose for writing this article was to make a permanent record of
the fact that what we had started was in fact completed and that all the
key components necessary for Linux to become a mature UNIX replacement
product were in the release of United Linux 1.0.
Prior to finalization of the specifications for United Linux 1.0, I met
with the Unix Systems Labs folks in Murray Hill, NJ, to validate that
the features targeted for United Linux 1.0 would not leave the product
short of its ability to replace UNIX. We all recognized that it would
take a few short years to harness full support for use in truly mission
critical applications, but the intent and purpose of Caldera being
behind United Linux was very well communicated and very clear internally
from the very beginning.
There was resistance from some of the old UNIX hands, but by and large
we all understood that the long-term ability of the company to sustain
UNIX was not financially and technically feasible. The SCO UNIX code
was all 32-bit with no 64-bit development pathway. The future of Linux,
with full capability to sustain mission-critical operation was
recognized to depend on 64-bit support and total robustness. Part of
that robustness requirement dictated the need to be able to run legacy
32-but UNIX applications under United Linux.
The inclusion of the Linux-ABI code and the ability to run OpenServer
binaries under United Linux were seen as essential from the outset.
Much more can be said here, but I will leave that discussion for a
Feel free to quote from this email.
John H Terpstra