| Revision History | ||
|---|---|---|
| Revision 1.3 | 2003-08-28 | rob |
| More updates from Craig Trader, Ed Hynan, Tony Taylor, Lon Jones, Vsevolod Sipakov, Earl Barr, Mark Taylor, Robert J. Harley, Eric Lesage, Andries Brouwer, Dick Porter, and dozens of other people who found the same things a bit later. (And me digging up yet more fresh material, and fixing some of the fresh material I dug up in 1.2.) | ||
| Revision 1.2 | 2003-08-26 | rob |
| Updates from Lars Nygard, Leon Brooks, Behrens Matt, Alastair Mayer, and a bunch of material from David Wheeler. Plus me getting kicked in the butt by the above to polish and fill in some bits. (And typos, typos, typos...) | ||
| Revision 1.1 | 2003-08-26 | esr |
| First round of minor corrections from the net. | ||
| Revision 1.0 | 2003-08-24 | rob |
| First published version. | ||
| Revision 0.1 | 2003-06-28 | esr |
| Document begun. | ||
Table of Contents
The amended SCO complaint against IBM filed on 16 June 2003 is, like its predecessor, a tissue of lies, deliberate distortions, and flimflam. This complaint strongly suggests that SCO has no real case, since it contains so many false and misleading statements. Instead, it appears that SCO is trying to turn worthless products into money by running up the stock price on false claims and/or trying to get bought out so the frivolous legal actions will stop.
Unlike its predecessor, this amended complaint has been brought to you by the generosity of Microsoft, who (on the evidence of SCO's 10-Q SEC filings) dropped at least six million dollars on SCO (plus a promise of five million more over the next three quarters) to help it make trouble for Linux.
SCO, having willingly made itself a sock puppet for the boys in Redmond, therefore becomes the first company other than Microsoft to have its utterances admitted to the gallery of infamy that is the Halloween Documents.
There follows the usual point-by-point takedown. Unlike SCO's claims, this analysis is based entirely on public information which third parties may verify by chasing links or through their local library.
Text in fixed-width font is verbatim from the complaint. Where our commentary is interspersed with material from the original complaint, our commentary is indented and highlighted in green as well. Key words and particularly egregious falsehoods in the original are highlighted in red.
We've also added some appendixes at the end to explain the history of this lawsuit: The Decline and Fall of Caldera, Sizing the Unix Market, Moore's Law (and how it applies here), and The Itanic Disaster.
______________________________________________________________________________
Brent O. Hatch (5715)
Mark F. James (5295)
HATCH, JAMES & DODGE, P.C.
10 West Broadway, Suite 400
Salt Lake City, Utah 84101
Telephone: (801) 363-6363
Facsimile: (801) 363-6666
David Boies
BOIES, SCHILLER & FLEXNER LLP
333 Main Street
Armonk, New York 10504
Telephone: (914) 749-8200
Facsimile: (914) 749-8300
Stephen N. Zack
Mark J. Heise
BOIES, SCHILLER & FLEXNER LLP
100 Southeast Second Street
Suite 2800
Miami, Florida 33131
Telephone: (305) 539-8400
Facsimile: (305) 539-1307
Attorneys for Plaintiff The SCO Group, Inc.
______________________________________________________________________________
IN THE UNITED STATES DISTRICT COURT
DISTRICT OF UTAH
______________________________________________________________________________
|--------------------------------------+--------------------------------------|
| | |
| THE SCO GROUP, INC., | |
| a Delaware corporation, | |
| | AMENDED COMPLAINT |
| Plaintiff, | |
| vs. | (Jury Trial Demanded) |
| | |
| INTERNATIONAL BUSINESS MACHINES | |
| CORPORATION, a New York corporation, | Case No. 03-CV-0294 |
| | |
| Defendant. | Hon: Dale A. Kimball |
|--------------------------------------+--------------------------------------|
_____________________________________________________________________________
Plaintiff, The SCO Group, Inc., a Delaware corporation, f/k/a Caldera Systems, Inc. (“SCO”), sues Defendant International Business Machines Corporation (“IBM”) and alleges as follows:
The plaintiff in this case, "SCO Group Inc.", was originally the Linux vendor "Caldera Systems Inc." It changed its name to "Caldera International Inc." on October 11, 2000 after purchasing the server division of The Santa Cruz Operation, and changed its name again to "SCO Group Inc." on May 15, 2003. The original SCO changed its name to "Tarantella Inc" on February 12, 2001 after divesting itself of its deteriorating Unix business to focus on its Tarantella application.
These are two distinct companies, one founded in 1979 and headquartered in California, and the other founded in 1994 and headquartered in Utah, which the plaintiff refers to interchangeably as SCO. To distinguish them, we refer to the plaintiff as SCO/Caldera and the older company the plaintiff purchased a server division from as SCO/Tarantella.
For the sad tale of Caldera's degeneration from an open-source pioneer to a parasitic lawsuit factory, see Appendix A.
1. UNIX is a computer operating system program and related software originally developed by AT&T Bell Laboratories (“AT&T”). UNIX is widely used in the corporate, or “enterprise,” computing environment.
Actually, ‘UNIX’ is not “a computer operating system program”. There is no single operating system you can point to and say "that alone is Unix, everything else is not Unix". Technically, the name "Unix" is a trademark of The Open Group describing a group of operating systems conforming to The Open Group's ‘Single Unix Specification’. More generally, Unix is a generic technical term used to describe a family of over a hundred operating systems, some of which meet The Open Group's trademark standard but many of which do not. Compaq's Tru64 was submitted and qualified for the Unix trademark, while Apple uses Unix in a generic sense when referring to its MacOS X. Microsoft's Windows NT has hosted a Unix compatibility environment called cygwin since 1998, allowing it to run Unix software.
While all Unixes in either sense share common ideas and methods, the ideas and methods they share are (and for twenty years have been) part of the common state of engineering art in the construction of operating systems, extensively documented in the public literature, and shared by a global community of hundreds of thousands of programmers and software engineers. The first paragraph of this complaint is the beginning of a deliberate, mendacious attempt to misrepresent this entire body of work as SCO's property.
SCO does not even own the Unix trademark, and they cannot be be ignorant of The Open Group, since both SCO and IBM are members. IBM is one of five "Sponsors", and SCO is one of 138 "Regular Members" — IBM's membership in The Open Group significantly outranks SCO's.
SCO itself does not offer a product named just ‘Unix’, it offers UnixWare (named after Novell Netware) and SCO OpenServer (which was once Microsoft Xenix).
Over the past 30 years, the basic ideas of Unix have permeated every aspect of the software industry. In 1996, Microsoft CEO Bill Gates gave a speech at Unix Expo where he called Windows NT "a form of Unix". The same "UNIX 95" certification SCO obtained for UnixWare was also granted to IBM's OS/390 mainframe operating system. A decade ago, IBM's OS/2 operating system for the PS/2 had an open source Unix emulation package called EMX (which was not created by IBM, but by a german individual named Eberhard Mattes).
Over the past 30 years, the basic ideas of Unix have permeated every aspect of the software industry. In 1996, Microsoft CEO Bill Gates gave a speech at Unix Expo where he called Windows NT "a form of Unix".
2. Through a series of corporate acquisitions, SCO presently owns all right, title and interest in and to UNIX and UnixWare operating system source code, software and sublicensing agreements, together with copyrights, additional licensing rights in and to UNIX and UnixWare, and claims against all parties breaching such agreements. Through agreements with UNIX vendors, SCO controls the right of all UNIX vendors to use and distribute UNIX. These restrictions on the use and distribution of UNIX are designed to protect the economic value of UNIX.
The “series of corporate acquisitions” draws a kindly veil over the fact that at least three different companies — AT&T, Novell, and the original SCO (now Tarantella) — have each discarded the rights to the increasingly obsolete System V codebase. Along the way, the intellectual property was significantly impaired.
The most obvious impairment of SCO's “right and title” to Unix source code and related IP is the 1993 settlement with the University of California at Berkeley, which legally confirmed the plurity of Unix by severing the BSD source base, despite its origin as a derivative work of AT&T code, from the control of AT&T's successors in interest.
USL's willingness to settle the lawsuit, "Unix System Laboratories v. Berkeley Software Design, Inc." (1993 U.S. Dist. LEXIS 19503; 27 U.S.P.Q.2D (BNA) 1721.), was motivated by the judge's ruling against issuing a preliminary injunction, which read in part:
Since Plaintiff has failed to provide enough evidence to establish a "reasonable probability" that Net2 or BSD/386 contain trade secrets, I find that Plaintiff has failed to demonstrate a likelihood of success on the merits of its claim for misappropriation of trade secrets. No preliminary injunction will issue.
The settlement resulted in the release of 4.3-Lite BSD, an open-source version of Unix explicitly free of any intellectual property claims from AT&T or its successors. The existence of unimpaired BSD imposed a massive new burden on any intellectual property claims putatively based on ownership of the System V code: to prove that any disputed IP could not have come from BSD.
Another significant source of impairment is the plethora of existing "perpetual" Unix licenses. These licenses were granted in return for a one-time payment requiring no ongoing royalties; they never expire, and they allow unlimited sublicensing and the creation of derivative works that explicitly belong to the licensee. IBM's original license to System V technology from AT&T is such a license, as shown by language in "Exhibit C" on SCO's own website, which is an amendment/clarification of IBM's System V contract. Paragraph 2 of Exhibit C explicitly allows IBM to create derivative works, and affirms that those derivative works belong entirely to IBM (except for any actual System V code snippets included in them, if any).
Regarding Section 2.01, we agree that modifications and derivative works prepared by or for you are owned by you. However, ownership of any portion or portions of SOFTWARE PRODUCTS included in any such modification or derivative work remains with us.
Although IBM's perpetual System V code license was with AT&T, SCO is still selling these perpetual licenses as the core offering of their SCOsource initiative. The second full paragraph on page 28 of their 2003 first quarter SEC filing begins:
Our SCOsource licensing revenue to date has been generated from license agreements that are non-exclusive, perpetual, royalty-free, paid up licenses to utilize our UNIX source code, including the right to sublicense.
A third source of impairment is the simple fact that tens of thousands of copies of Unix source code have been circulating within the community for three decades. SCO itself puts the number at over 30,000 licensing agreements with 6000 entities (from the same 2003 Q1 SEC filing, bottom of page 20). Maintaining trade secret status on something like that is effectively impossible. One of the authors of this commentary is collecting evidence from senior Unix gurus who have source code to System V and derivatives in their archives, to prove that this code cannot have trade secret status.
It is worth noting that when SCO says “These restrictions on the use and distribution of UNIX are designed to protect the economic value of UNIX.” it is using the phrase “economic value” in a tendentious way. No action of IBM has altered the economic value of Unix; a copy of Unix has much the same effect on a user's ability to process data and do things as it did before any of IBM's actions. The price that can be charged for that value, the amount of rent SCO can collect, has been diminishing for years, but this is nothing new. SCO's own SEC filings have proposed a number of alternative hypotheses for their poor economic performance, and before 2003 SCO/Caldera had never had a single profitable quarter. They're not complaining about the loss of actual revenue, but about the loss of a purely hypothetical market opportunity that might conceivably have resulted in revenue.
3. A variant or clone of UNIX currently exists in the computer marketplace called “Linux.” Linux is, in material part, based upon UNIX source code and methods, particularly as related to enterprise computing methods found in Linux 2.4.x releases and the current development kernel, Linux 2.5.x. Significantly, Linux is distributed without a licensing fee and without proprietary rights of ownership or confidentiality.
Referring to Linux as a “clone” presumes the conclusion SCO wishes the court to reach. Linux is not in essence a clone or derivative work of any SCO property, but an autonomous line of evolution of generic, publicly documented Unix ideas, with a history of independent development documented back to its beginnings. Even a showing that Linux incidentally or accidentally contained some Unix code would not alter Linux's essential nature as an independent work.
Linus Torvalds explained the situation in the chapter he wrote for the 1999 book "Open Sources" (the complete text of which is available online):
Linux is a Unix-like operating system, but not a version of Unix. This gives Linux a different heritage than, for example, Free BSD. What I mean is this: the creators of Free BSD started with the source code to Berkeley Unix, and their kernel is directly descended from that source code. So Free BSD is a version of Unix; it's in the Unix family tree. Linux, on the other hand, aims to provide an interface that is compatible with Unix, but the kernel was written from scratch, without reference to Unix source code. So Linux itself is not a port of Unix. It's a new operating system.
A quick summary of early Linux development circa 1991 and 1992 is available from Linux International. (Notice that the first documentary evidence of the Linux project on the internet was Linus Torvalds asking for a copy of the Posix standard, not for source code.) The Linux development process has always been public; an archive of the "linux-activists" mailing list dating back to october 1991 is available online. The book "Just for Fun", by Linus Torvalds and David Diamond, covers the early days of Linux in detail. Linus created Linux shortly after taking a course on operating system design from the University of Helsinki, which used the textbook Operating Systems: Design and Implementation by Dr. Andrew Tanenbaum, which was about a tiny unix-like operating system called "Minix" which Tanenbaum had written from scratch as a teaching tool. [1] Linus chose a very different design from Minix, and used no Minix code in Linux, even though Minix itself was an independent work of Tanenbaum's which contained no System V code. As Linux became more popular, Tanenbaum publicly disagreed with Linus' design choices, and a subsequent public debate took place on the internet in January 1992 between Tanenbaum and Torvalds on operating system design principles. The complete source code of many Linux kernel versions dating back to 1991 is freely downloadable from www.kernel.org; you can download the very first release (or view the release notes), the first semi-stable version (release notes), or the 1.0 release. Linus Torvalds has been interviewed numerous times, here are a 1992 interview and a 1994 interview. Here is an academic analysis of the history of Linux kernel development. Here is a recent weekly technical summary of Linux-kernel development activity, which should give even a layman an example of the openly documented nature of the development, and may be considered a representative example from the archive of such technical summaries going back to the beginning of 1999.
Historian Peter Salus, author of "A Quarter Century of Unix" and executive director of Usenix for three years back in the 1980s, gave a talk about the History of Linux for the tenth anniversary of Linux, a recording of which is available online and as a downloadable MP3 (1 hour and 11 minutes). (He also gave an excellent talk at Usenix in 2001 on A History of Unix and Unix Licenses (50 minutes).)
Throughout this complaint, SCO attempts to confuse the widely admitted sharing of ideas and methods among computer professionals and Unix-family operating systems with putatively illicit misappropriation of source code. This attempt begins here with the introduction of the term “methods” But the history is clear; Linux could be “based on” Unix methods through published treatment such as Maurice Bach's The Design of the Unix Operating System, through the Berkeley Unix line of source code (also well documented in academic textbooks such as The Design and Implementation of the 4.4 BSD Operating System), or in several other ways that would not imply any legal connection to or proprietary entitlement by AT&T's successors in interest.
During the months since its initial complaint was filed, SCO has repeatedly refused to make public any actual evidence that Linux is “in material part, based upon UNIX source code”, preferring instead to trumpet its claims in the media. (Their one gesture in the direction of disclosure, at SCOforum in August 2003, proved to be of code they don't own.) In any case, given that the revision history of the SCO source code is not (unlike that of Linux) a matter of public record readily accessible on numerous archives, the burden should be on SCO to show that any similarities have not been covertly and deliberately inserted by SCO itself for the purpose of enabling SCO to commit barratry against IBM and the Linux community at large.
4. The UNIX software distribution vendors, such as IBM, are contractually and legally prohibited from giving away or disclosing proprietary UNIX source code and methods for external business purposes, such as contributions to the Linux community or otherwise using UNIX for the benefit of others. This prohibition extends to derivative work products that are modifications of, or based on, UNIX System V source code or technology. IBM and certain other UNIX software distributors are violating this prohibition, en masse, as though no prohibition or proprietary restrictions exist at all with respect to the UNIX technology. As a result of IBM's wholesale disregard of its contractual and legal obligations to SCO, Linux 2.4.x and the development Linux kernel, 2.5.x, are filled with UNIX source code, derivative works and methods. As such, Linux 2.4.x and Linux 2.5.x are unauthorized derivatives of UNIX System V.
This paragraph begins with a sleight of hand. By using the phrase “proprietary Unix source code and methods” in a way that leaves it ambiguous whether “proprietary” is a restrictive modifier or not, SCO attempts to confuse a trivial but true claim (disclosure of proprietary methods is forbidden) with an extremely broad and false claim (disclosure of Unix methods is forbidden).
But, in fact, it has not been shown that there are any proprietary methods or any proprietary source code at issue. SCO holds no Unix patents; the state and disposition of the Unix copyrights is unclear and presently disputed between SCO and Novell; finally, by making the source code “readily ascertainable by proper means” including the sale of source licenses to educational institutions, SCO and various predecessors in interest have failed to meet the Utah standards for maintenance of trade secrecy in the Unix source code. Their "trade secrets" have been taught in college classes for 20 years, with their knowledge and approval.
On August 18 2003 at the annual SCOforum show, SCO tried to back up its contention that Linux is “filled with UNIX source code, derivative works and methods” — and was promptly shot down.
As for derivative works, SCO needs to read its own Exhibit C, paragraph 2, which states “...modifications and derivative works prepared by or for you are owned by you.”. (The “you” in this case being IBM.) More on this in the response to paragraph 6.
As a final point, the moment that SCO shipped a Linux kernel containing any included System V code, it knowingly put that code under the terms of the GNU General Public License, and that code thereby ceased to be proprietary (and if SCO/Caldera did not accept the GPL, then they willfully violated the copyrights of every single Linux contributor). SCO and predecessor in interest Caldera shipped Linux source code for eight years, and joined numerous Linux development consortia including The Trillian Project and United Linux. Although SCO tried to cover its tracks by suspending distribution of its own Linux product two months after filing its lawsuit against IBM, Linux source code is still available for download from SCO, and must remain so for three years according to the terms of secton 3 of the GNU General Public License under which Linux was made available to SCO/Caldera for distribution. SCO/Caldera complied with its obligations under the GPL for eight years, and can hardly start claiming it was unaware of the terms of that license, since they submitted a complete copy of the GPL to the securities and exchange commission (attached as exhibit 10.14 to their S-1 Inital Public Offering filing [2]). The GPL was also the contract under which they sold Linux to their customers, thereby binding themselves to an obligation to those customers.
SCO voluntarily chose to participate in the Linux market. It was never required by any third party to do so, and found doing so profitable (in gross terms, if not net). If SCO didn't exert due diligence on the code SCO itself shipped, it can't blame IBM. In fact, Linux was SCO/Caldera's core business before it purchased the archaic Unix line from SCO/Tarantella three years ago. When Caldera first acquired SCO, its SEC filings stated the intention to “unify” the Linux and Unix lines, specifically questions 29 and 31 of the question and answer session:
29. How will the products of both companies be integrated?
We will integrate all products as appropriate based on Caldera, Inc.'s focus of common application, development and management.
31. How will this affect the Linux industry/community?
Caldera, Inc. intends to unify the technologies of Linux and UNIX enhancing the attractiveness to Linux and business customers. In addition, Caldera, Inc. will provide open access to those technologies, greatly benefiting the Linux community and industry.
Finally, this paragraph concludes with an attempt to parlay a narrow claim of IBM contract violations into a blatant land grab against all current versions of Linux, including millions of lines of code which are not derived from System V code by any standard, however inclusive. This is overreaching of the crudest sort.
5. As set forth in more detail below, IBM has breached its obligations to SCO, induced and encouraged others to breach their obligations to SCO, interfered with SCO's business, and engaged in unfair competition with SCO, including by:
a) misusing UNIX software licensed by SCO to IBM and Sequent;
b) inducing, encouraging, and enabling others to misuse and misappropriate SCO's proprietary software; and
c) incorporating (and inducing, encouraging, and enabling others to incorporate) SCO's proprietary software into Linux open-source software offerings.
In this summary of claims SCO speaks only of software, not of methods or algorithms. The implied breach is of SCO's copyrights (if any). Elsewhere, however, SCO speaks of methods. The effect is to deliberately confuse two issues and prejudice claims which SCO could not otherwise establish.
6. As a result of these breaches, SCO sent a notice of termination to Mr. Sam Palmisano, the Chief Executive Officer of IBM on March 6, 2003. The termination notice specified that, pursuant to SCO's contractual rights under controlling agreements, IBM's right to use or distribute any software product based on UNIX System V technology, including its own version of UNIX known as “AIX,” would be terminated on June 13, 2003, unless such breaches were reasonably cured prior to that time.
IBM has pointed to the plain language of its license agreement as a refutation of any claim of breach. This language, from Exhibit C, includes the following in paragraph 9:
Nothing in this agreement shall prevent LICENSEE from developing or marketing products or services employing ideas, concepts, know-how, or techniques relating to data processing embodied in SOFTWARE PRODUCTS subject to this agreement, provided that LICENSEE shall not copy any code from such SOFTWARE PRODUCTS into any such product [...] If information relating to a SOFTWARE PRODUCT subject to this agreement at any time becomes available without restriction to the general public by acts not attributable to LICENSEE or its employees, LICENSEE's obligations under this section shall not apply to such information after such time.—
Since SCO itself ensured that any System V code in Linux would become available without restriction to the general public by shipping it in a Linux kernel for eight years (up through version 2.4.14), the conditions of this language have been met. In addition, the source code for Unix System III, an earlier version of Unix containing a significant subset of the later System V code, was made available for free download on SCO's website under a click-through "Ancient Products License". And the BSD lawsuit confirmed the release of a significant portion of the methods, if not the actual code, of System V Unix.
7. The termination notice was based, in part, on IBM's self-proclaimed contributions of AIX source code to Linux, and use of UNIX/AIX methods for accelerating the development of Linux in contravention of IBM's contractual obligations to SCO.
IBM's license entitles it to use Unix concepts and methods, to make derivative works, and to sublicense Unix source code and technologies to others.
If SCO is going to try to quote IBM's marketing people out of context, the obvious response is to quote SCO's CEO in context. The CEO of Caldera during the SCO acquisition and for a number of years afterwards was Ransom Love. An interview Ransom Love gave in August 2001 (a year after the SCO acquisition) includes several interesting statements. In answer to the question “What is your position [on open source] now?”, Love replied “Caldera is absolutely committed to open source... We are going to continue publishing under open source.” In answer to the question "What does the future hold for your unified Linux/Unix platform?", Love replied "With UnixWare we can now take Linux to 32-way systems... So we are taking both and combining them into one platform." When asked about OpenServer, Love said "...what we plan to do is to take the OpenServer technology to Linux, probably with some sort of open source source licence..." When asked "What does the recent acquisition of SCO really mean for Caldera?", Love replied "Our mission is to enable development, deployment and management of a unified Linux and Unix operating system."
If the current management of SCO is completely unaware of what their CEO claimed the company's mission was two years ago, they are incompetent. If they are aware of it, they're being blatantly dishonest. SCO/Caldera devoted entire securities and exchange commission filings to unifying Linux and Unix.
8. Pursuant to its rights under the controlling agreements, IBM was entitled to 100 days to cure its underlying contractual breaches, provided it was willing and able to do so. Both parties were contractually required to “exert their mutual good faith best efforts to resolve any alleged breach short of termination.”
Nowhere does SCO specify exactly what actions SCO wanted IBM to undertake. What, specifically, did SCO want IBM to do? It can't have wanted IBM to remove offending code from the Linux Kernel, because it refuses to reveal to IBM or the Linux community which code it considers to be infringing. SCO CEO Darl McBride explained his reason for refusing to identify allegedly infringing code in the Linux kernel as follows: "The Linux community would have me publish it now, (so they can have it) laundered by the time we can get to a court hearing. That's not the way we're going to go." SCO actively does not want IBM or other Linux participants to stop infringing, and has intentionally made it impossible for them to comply with SCO's stated wishes. This cannot be considered a good faith best effort on SCO's part.
9. To that end, SCO did everything reasonably in its power to exert a good faith effort to resolve the termination of IBM's UNIX contract rights. Conversely, during the 100-day period, IBM did not set forth a single proposal or idea for cure.
SCO's phrasing here is particularly misleading. Notice they don't say they made any effort to avoid the termination of IBM's contract. In this context, “resolve” means “enact”, or “hasten”. The contract language (again in Exhibit C, paragraph 5 this time) forced them to wait 100 days. It's too bad that SCO's promise not to terminate for “breaches we consider to be immaterial” didn't require arbitration by a disinterested third party. Asking SCO to judge what is and isn't immaterial is like asking the fox to guard your chickens.
As for the rest of the paragraph, IBM has maintained that there is no breach. Where there is no breach, there is no need for cure.
10. SCO has therefore terminated IBM's right to use any part of the UNIX System V source code, including its derivative AIX, effective as of June 13, 2003 (the “AIX Termination Date”).
11. As of the AIX Termination Date, IBM is contractually obligated to discontinue use of and return or destroy any and all copies of the Software Products defined in the controlling agreements, which include UNIX System V source code and all its derivatives, including AIX.
As a matter of public record, IBM has maintained that its license is perpetual and irrevocable. This position is based on the language of Amendment X, dated 1996.
It's also interesting to note that that the “perpetual” licenses that SCO is offering through SCOsource are the same general kind of license: that is, they seem only to be guaranteed to last until SCO needs money.
12. Plaintiff SCO is a Delaware corporation with its principal place of business in Utah County, State of Utah.
13. Defendant IBM is a New York corporation with its principal place of business in the State of New York.
This case was removed from Utah state to Federal court because of errors in SCO's original complaint. One of these was giving IBM's place of business incorrectly. Nice to see they got it right this time.
14. Sequent Computer Systems, Inc. (“Sequent”) was formerly an Oregon corporation that contracted with SCO's predecessor in interest, AT&T. Sequent was subsequently merged into IBM in a stock transaction.
Sequent started in 1983 when a group of engineers left Intel to work on parallel computing. Their first product, released in 1984, ran a modified version of BSD called Dynix. Their first Intel 80386 based product shipped commercially in 1987. Like many in the industry, Sequent shifted from BSD to System V in response to the uncertain legal status of the BSD code before the USL vs BSD lawsuit was resolved. The (theoretically) System V based product was called Dynix/ptx. In reality, Dynix/ptx was still based on BSD, with some System V interface wrappers so it could claim System V compatability, and an AT&T license because the USL vs Berkeley lawsuit hadn't been resolved in Berkeley's favor yet. (Wikipedia has an excellent article on Sequent's history.)
15. This Court has subject matter jurisdiction pursuant to 28 U.S.C. §1332 in that diversity of citizenship exists between the parties and the amount in controversy exceeds $75,000, exclusive of interest and costs.
16. This Court has in personam jurisdiction over IBM pursuant to Utah Code Ann. §78-27-24 on the bases that IBM is (a) transacting business within this State, (b) contracting to provide goods and services within this State and (c) causing tortious injury and breach of contract within this State.
Point c. being the claim they hope to prove, not a statement of fact.
17. Venue is properly situated in this District pursuant to 28 U.S.C. §1391 in that IBM maintains a general business office in this District and a substantial part of the events giving rise to the claims alleged herein occurred in this District.
If SCO's claims had any merit, the latter assertion would almost certainly be false. IBM's largest Linux development center in the continental US is the Linux Technology Center in Austin, Texas. SCO's corporate headquarters during the Monterey project (the period of the disputed contract) was in California, with a large developer concentration in Canada. Since the "events giving rise to the claims alleged herein" occurred largely in the fevered imagination of various SCO executives and nowhere else, a case could be made that that they took place in Utah. The reality is that Utah is the venue only because SCO wanted a hometown advantage, and it's dishonest to claim otherwise. (This is a trivial error in comparison to those in the rest of their complaint.)
18. UNIX is a computer software operating system. Operating systems serve as the link between computer hardware and the various software programs (“applications”) that run on the computer. Operating systems allow multiple software programs to run at the same time and generally function as a “traffic control” system for the different software programs that run on a computer.
This claim is technically defective. Not all operating systems allow multiple programs to run at the same time.
19. By way of example, in the personal computing market, Microsoft Windows is the best-known operating system. The Windows operating system was designed to operate on computer processors (“chips”) built by Intel. Thus, Windows serves as the link between Intel-based processors and the various software applications that run on personal computers.
20. In the business computing environment for the Fortune 1000 and other large corporations (often called the “enterprise” environment), UNIX is widely used.
However, SCO's own versions of Unix have never had significant enterprise market share. The technology to support that use had to be supplied by others, as we document in the first OSI Position Paper on the SCO-vs.-IBM Complaint.
In terms of revenue, the top three players in 1999 were Sun (38 percent of the market), Hewlett Packard (24 percent), and IBM (16 percent), together soaking up a combined total of 78 percent of the Unix market's revenue and leaving SCO/Tarantella part of a long list of financial also-rans. SCO's market share came by pricing their systems so low they wound up in financial difficulty and sold their Unix business to Caldera, to finance a transition to something they considered potentially more profitable.
To answer the question of exactly how widely SCO's own Unix products are used, it's necessary to point out that SCO sells two very different versions of Unix. SCO OpenServer is a low-end product that traces its heritage back to Microsoft Xenix in 1979. OpenServer is primarily used for point of sale systems such as fast food cash registers. SCO's other Unix line is UnixWare, which was named after Novell NetWare and is the Unix product SCO acquired from Novell in 1995. SCO put three years of intense development effort into UnixWare between 1995 and 1998 in an attempt to make an enterprise operating system out of it, but never managed to achieve any significant marketshare or market interest.
SCO's low-end product, OpenServer, provides the majority of their product revenue and the vast majority of their unit volume. According to page 22 of SCO's first quarter 2003 SEC filing::
SCO's first quarter 2003 Company's products revenue.
For the second quarter of fiscal year 2003, OpenServer revenue was $5,750,000, or 52 percent of total products revenue and UnixWare revenue was $3,017,000, or 27 percent of total products revenue. For the first two quarters of fiscal year 2003, OpenServer revenue was $10,491,000, or 47 percent of total products revenue and UnixWare revenue was $7,181,000, or 32 percent of total products revenue.
SCO is not a significant "enterprise Unix" player not only because their entire quarterly revenue from their "enterprise" product (UnixWare) is less than 1/10th that of Sun, IBM, or HP (which is not a new development), but because their low-end product provides the majority of their products revenue, and only on a single hardware platform. They are primarily a niche player in the low-end PC Unix market.
It's also interesting to note that any special heritage from AT&T that SCO claims to have acquired through Novell would have been embodied in UnixWare, yet UnixWare is not SCO's best-selling product. And OpenServer, which provides the majority of their revenue today, was at one point a discontinued piece of trash that Caldera wanted to leave with SCO/Tarantella when it first tried to purchase SCO's operating systems division, as explained in the February 9, 2001 article, "Caldera Systems expands Unix acquisition plans":
Under the previous terms of the acquisition, announced in August, Linux seller Caldera Systems planned to acquire SCO's UnixWare product. Now it plans to also acquire SCO's earlier version of Unix, OpenServer, Caldera said in a statement Friday.
SCO stopped development of OpenServer in 1999, and Caldera Systems plans to continue with SCO's practice of only minor updates while encouraging customers to migrate to other operating systems, Tamang said.
OpenServer remains SCO's cash cow (in gross, if not net terms) because they were unsuccessful in their efforts to migrate customers off of it, just as they were unsuccessful in their efforts to repurpose the acquired reseller network to sell Linux instead of obsolete proprietary versions of Unix.
21. The UNIX operating system was originally built by Dennis Ritchie, Ken Thompson and other software engineers at AT&T Bell Laboratories. After successful in-house use of the UNIX software, AT&T began to license UNIX as a commercial product for use in enterprise applications by other large companies.
This paragraph ignores huge contributions by many developers outside of AT&T, many of whom are active in the Linux community today. Specifically, at the time Unix was invented at Bell Labs, AT&T was forbidden from commercializing it by a Jan 24, 1956 antitrust consent decree[3]. Specifically, "An injunction was issued which barred AT&T from engaging in any business other than the provision of common carrier communication services, and required Western Electric and AT&T to license their patents to anyone who wanted them upon the payment of appropriate royalties."
Since AT&T was forbidden from entering the computer business, they could not sell Unix as a product. The outside world found out about Unix after Ken and Dennis presented a paper on their Unix research at a 1973 ACM operating systems conference, and the paper was widely published to the academic community in the proceedings of that conference in 1974. The 1956 consent decree required AT&T to license its technology to outsiders who asked, but their offical policy concerning these academic Unix licensees was: "No advertising, no support, no bug fixes, payment in advance."
Unix could not become a commercial product of AT&T before the 1984 divestiture undid the 1956 antitrust injunction keeping them out of the computer business. During this decade-long gap, there was no commercial incentive for AT&T to invest in Unix development, but significant incentive for early open-source development to occur, largely co-ordinated by the University of California at Berkeley after Ken Thompson's teaching stint there (see the response to Paragraph 3).
Ignoring these contributors is crucial to SCO's case; were it to acknowledge them, it would have to admit to the substantial impairment of its rights by the settlement of the 1993 BSD case, which hinged on AT&T's misappropriation of work done at Berkeley and elsewhere. See, again, the OSI Position Paper on the SCO-vs.-IBM Complaint.
22. Over the years, AT&T Technologies, Inc., a wholly owned subsidiary of AT&T, and its related companies licensed UNIX for widespread enterprise use. IBM, Hewlett-Packard, Inc. (“HP”), Sun Microsystems, Inc. (“Sun”), Silicon Graphics, Inc. (“SGI”) and Sequent became some of the principal United States-based UNIX licensees, among many others.
It is noteworthy that they do not mention themselves, the original Santa Cruz Operation, SCO/Caldera's predecessor in interest. SCO sublicensed Xenix from Microsoft, and in 1987 Microsoft divested itself of Xenix to focus on DOS, OS/2, and Windows. Xenix went on to become SCO's primary product, and remains so today under the name SCO OpenServer. In the 1980s, SCO was one of many licensees of AT&T System V Unix. Perhaps they belong in the "among many others" category because they were not an enterprise player.
Microsoft's CEO Bill Gates reminisced about Microsoft Xenix in a 1996 speech at Unix Expo, where he was trying to recruit Unix developers to port their applications to Windows NT:
If we go way back in time, Microsoft was actually the first one to go to AT&T and beg to get a nice high-volume commercial license for Unix. And for many, many years we were the highest volume licensee, not only for our own Xenix products, but Siemens with theirs, Santa Cruz with theirs, and dozens and dozens of sub-licensees.
I have to admit, it was fairly difficult to work with AT&T back then. They simply didn't understand what they had. They didn't understand how to manage the asset, either in terms of promoting it properly or in terms of making sure that there wasn't fragmentation in how different implementations were put together. And so that vacuum in leadership created a bit of a dilemma for everybody who was involved in Unix.
Well, Microsoft stepped back and looked at that situation and said that the best thing for us might be to start from scratch: build a new system, focus on having a lot of the great things about Unix, a lot of the great things about Windows, and also being a file-sharing server that would have the same kind of performance that, up until that point, had been unique to Novell's Netware.
And through Windows NT, you can see it throughout the design. In a weak sense, it is a form of Unix. There are so many of the design decisions that have been influenced by that environment. And that's no accident. I mean, we knew that Unix operability would be very important and we knew that the largest body of programmers that we'd want to draw on in building Windows NT applications would certainly come from the Unix base.
23. IBM, HP, Sun, SGI and the other major UNIX vendors each modified UNIX to operate on their own processors. Thus, for example, the operating system known as “HP-UX” is HP's version of UNIX. HP-UX is a modification of and derivative work based on UNIX System V source code.
And Microsoft modified Unix to operate on cheap 16 bit 8086 PCs which were primarily used for running DOS, and sublicensed it to the Santa Cruz Operation.
24. Similarly, the operating system known as Solaris is Sun's version of UNIX. Solaris is a modification of, and derivative work based on, UNIX System V source code.
Solaris is the second major Unix operating system from Sun. Its first, SunOS, was based on BSD code rather than AT&T's version, as the founding of Sun predated the AT&T divestiture and one of Sun's four founders, Bill Joy, had been Berkeley's head of Unix development as a graduate student. After the divestiture, AT&T purchased 19 percent of Sun and convinced them to migrate from BSD to the System V codebase (hence from SunOS to Solaris). The real picture is far more complicated than SCO presents: Commercial Unix existed before AT&T commercialized its own Unix. When AT&T entered the Unix market after the divestiture, it found itself in competition with the existing Unix players, leading to "the Unix wars" and the spin-off of Unix System Labs (USL) from AT&T (and Novell's subsequent purchase of USL).
25. SGI's UNIX-based operating system is known as “IRIX.” IRIX is a modification of, and derivative work based on, UNIX System V source code.
26. IBM's UNIX-based operating system is known as “AIX.” AIX is a modification of, and derivative work based on, UNIX System V source code.
Again, SCO oversimplifies. According to page 26 of the book "Life with Unix" by Don Libes and Sandy Ressler, IBM entered the Unix market in 1983 with a version of Unix (for the 16-bit 8086 based IBM PC) sublicensed from a company called “Interactive”. This predated AIX by a number of years. In the late 1980s, IBM also offered another version of Unix as an alternative to its System V based AIX: the BSD-based Academic Operating System (AOS). AIX is still around while IBM's other Unix offerings have died out because over the years IBM invested significant amounts of internal development effort into keeping AIX relevant.
27. Sequent's UNIX-based operating system is known as “DYNIX/ptx.” DYNIX/ptx is a modification of, and derivative work based on, UNIX System V source code.
See the response to paragraph 14 for the history of Dynix.
28. The various identified versions of UNIX are sometimes referred to as UNIX “flavors.” All commercial UNIX “flavors” in use today are modifications of and derivative works based on the UNIX System V Technology (“System V Technology”). Were it not for UNIX System V, there would be no UNIX technology or derivative works available for IBM and others to copy into Linux.
This last is a serious and willful misrepresentation of the facts. It is historically true that many (though not all) modern Unixes derive methods and ideas from System V. But the 1993 settlement, which forbade Novell and all their successors in interest (including SCO) from asserting further IP claims against the Berkeley source code, established that mere derivation of methods is not sufficient to establish a proprietary claim on the code.
SCO's claim that no commercial Unix product exists independent of System V means that BSDi's successor in interest, Wind River, is also a target despite the 1993 settlement. [4]
SCO also deliberately ignores other lines of Unix development, like SunOS, which either predated System V or were based on BSD. The claim in paragraph 28 is simply false. It is factually untrue. Sun was quite profitably selling SunOS from its founding until AT&T purchased a significant equity stake in the company to convince it to change horses. IBM could easily have made a commercial success of AOS. The Santa Cruz Operation's founding in 1979, and the early history of its core product OpenServer/Xenix, predates the AT&T divestiture and corresponding release of System V as the first commercial software product from AT&T. Again, SCO wilfully ignores the impact of a decade of Berkeley led open-source development of BSD, because the truth undermines their case.
29. SCO is the sole and exclusive owner of all Software and Sublicensing Agreements that control use, distribution and sublicensing of UNIX System V and all modifications thereof and derivative works based thereon. SCO is also the sole and exclusive owner of copyrights related to UNIX System V source code and documentation and peripheral code and systems related thereto.
Both these claims are materially false for reasons already adduced, although this is the first time they mention copyright. Again, they are attempting to confuse issues; they are not actually making copyright claims.
The copyrights to System V may also have been impaired by the BSD lawsuit, the issue was left unresolved and would have to be reopened if System V copyright claims were to be pressed. AT&T included code from BSD, SunOS, Xenix, and possibly other sources in System V, but did not always get proper permission, and even stripped off third-party copyright notices from the code it incorporated. It assumed that its ownership of Unix gave it carte blanche ownership of derivative works, an assumption which came back to haunt it during the BSD lawsuit. The one thing SCO probably fears most is a judge taking the 1993 settlement as a precedent.
In addition, Novell, the company that owned System V before SCO/Tarantella, has stated claims that the copyrights did not pass to SCO but remain with Novell. SCO/Caldera claims to have come up with an amendment to the contract between SCO/Tarantella and Novell, which Novell itself cannot find a copy of in its files. This issue remains unresolved, and Novell continues to strongly object to SCO's complaints against IBM and the Linux community.
Even if SCO/Caldera's amendment proves to be genuine, SCO had not registered a transfer of copyright with the Patent and Trademark office at the time of the alleged contract violation, or even at the time of filing its first complaint against IBM in Utah court. As such, any damages SCO could potentially be awarded against IBM for violation of those unregistered copyrights are capped by statute at $150,000 per infringing work. Considering that Microsoft has already given SCO several million dollars to pursue the lawsuit against IBM and Linux, and they've already spent around two million dollars on legal fees, copyright damages are not significant.
30. During the 1990s the enterprise computing market for high-performance workstation computers came to be dominated by UNIX and the primary UNIX vendors identified above, each supplying its own version of the UNIX operating system based on UNIX System V Technology. UNIX became synonymous with “workstation” computers that typically operated on a RISC processing platform.
Sun became the leading vendor in the 1980s with SunOS, which was based on BSD. It switched to System V years later, motivated by a large equity investment from AT&T.
SCO/Tarantella was not a workstation vendor on a RISC platform. Until 1995 it exclusively sold low-end Unix systems for cheap PC hardware. Meanwhile SCO/Caldera was founded in 1994 as a commercial Linux vendor, also on cheap PC hardware, and had no involvement with System V before purchasing SCO's server division in early 2000.
Interestingly, AT&T directly sold "386 Unix System V" (in competition with SCO and other PC Unix vendors) in 1987, on 3.5" and 5.25" floppy disks. This was during AT&T's brief attempt to become a PC vendor (as documented in this May 1986 article from Byte Magazine reviewing one of their hardware offerings). AT&T was remarkably unsuccessful competing against both commodity PC hardware and the high-end workstation market led by established vendors like Sun, and quickly spun off Unix Labs as a seperate company, which was acquired by Novell.[5]
31. The RISC processing platform provides high-power computing capabilities at a relatively higher price for “workstation” computing. The alternative to “workstation” computing is commonly known as “desktop” computing on personal computers. The operating system market for “desktop” personal computers is dominated by Microsoft Corporation and its various Windows-based operating system products. The reason for this distinction is that most desktop computers (PC's) are designed to operate on Intel and Intel-compatible computing platforms. Most workstations are designed to operate on variants of RISC processing platforms and RISC-compatible computing platforms. PC systems and RISC systems are not compatible with each other. Thus, most versions of UNIX will not operate on Intel-based PC's for desktop computing; and Windows will not operate on RISC-based workstations for enterprise computing.
SCO glosses over the fact that IBM invented RISC technology. SCO also ignores the existence of PC ports of several Unix variants, such as IBM's AIX for the PS/2, or Sun's Solaris x86. In terms of Windows not operating on RISC systems, remember that Microsoft sold a version of Windows NT for the Alpha processor, a RISC system, for several years before deciding there wasn't enough money in it. (Versions of Windows NT 4.0, up until service pack 3, were also available for the PowerPC and Mips processors, also RISC systems.)
32. Most of the primary UNIX vendors identified above did not attempt to develop a UNIX “flavor” to operate on an Intel-based processor chip set. This is because the earlier Intel processors were considered to have inadequate processing power for use in the more demanding enterprise market applications.
Compare this language with that of the original complaint, in which SCO falsely claimed to have been the sole 386 vendor in this market. Of the vendors SCO has listed so far (IBM, Sun, HP, SGI, and Sequent), the majority did produce Intel versions. IBM had AIX for the PS/2, Sun still sells Solaris for x86 (and gave away copies for several years under its Free Solaris Binary License Program), and Sequent shipped its first 386 based NUMA system back in 1985. That's three of the five, which is "most".
The very title of this section implies a fallacy. The IBM PC was created by IBM. Shipping millions of units of hardware annually, which were powered by Intel processors, created a market niche for software vendors like SCO.
Furthermore, SCO's first Unix product, Xenix, was commissioned by Microsoft, the company behind the dominant PC operating system "Windows" as described in Paragraph 19 of this complaint. Microsoft was the AT&T licensee and the owner of the Xenix copyrights until 1987. SCO sublicensed Xenix from Microsoft, and was at the time one of several Xenix resellers for Microsoft. If SCO is claiming that the products it shipped created a market for Unix on IBM PC hardware, the credit would belong to Microsoft, not SCO.[6]
Also, the title is inaccurate: calling the x86 market the "Intel" market is like calling the PC market the "IBM" market. Just as the IBM PC defined a commodity niche for many similar computer systems, so has Intel's x86 line spawned a range of clones. Companies other than Intel currently active in the x86 processor market include Advanced Micro Devices (makers of the AMD Athlon, and important for reasons explained in Appendix D), National Semiconductor (big in laptops and low power devices), and Transmeta (Linus Torvalds' former employer). The Unix operating system family itself forms a third commodity niche, along with commodity PC systems running commodity x86 processors.
33. As computers grew in popularity to perform business functions, the processing power of Intel-based processor chips also began to increase dramatically. Consistent with Intel founder Gordon Moore's prediction, computer chips remained inexpensive while exponentially increasing in power and performance.
SCO doesn't bother to explain Moore's Law, which is important for understanding both why there was no special insight in SCO/Tarantella's business plan and why the System V technology SCO touts so highly is nearly valueless today.
Appendix C describes Moore's Law in more detail, along with several consequences of Moore's Law that SCO would rather avoid mentioning. Among the points covered in the appendix are that Moore's law offered no special insight to SCO/Tarantella that was not available to other vendors, that Moore's Law was not limited to Intel chips but applies to competing products (such as Motorola's 68k processor line, the main competitor to Intel chips in the 1980's), and that an inescapable consequence of Moore's Law is an annual depreciation rate for computer technology that renders System V technology nearly valueless a decade and a half after its introduction.
34. Seeing this emerging trend, it became evident to SCO that Intel chips would gradually gain widespread acceptance for use in the enterprise marketplace.
It was clear to all that microchips would eventually take over the enterprise market, due to Moore's Law. But the x86 line's dominance among microchip vendors was not the result of Moore's Law, which affected them all roughly equally, but due to IBM's endorsement of one specific microprocessor line (Intel's x86) for its PC, instead of Motorola's 68k, Zilog's Z8000, or other competing microprocessor lines. If IBM had gone with another processor for its PC, its descendants are what modern computers would be using today. No reasonable industry expert would dispute this.
35. Therefore, while other major UNIX vendors modified UNIX for their own respective RISC-based computing platforms, SCO developed and licensed the UNIX-based operating system for Intel-based processors for enterprise use that is now known as “SCO OpenServer.”
Here SCO/Caldera is explicitly claiming that SCO OpenServer, the successor to Xenix, was intended for enterprise use. Not UnixWare, but OpenServer, which had its development discontinued in 1999 under SCO/Tarantella due to irredeemable obsolecense. They seem unclear on the definition of "enterprise", which may be one reason they've never been successful in the enterprise market. OpenServer is used to run cash registers, not Fortune 500 data centers.
Background on the economic impact of Moore's Law is provided in Appendix C.
36. SCO's early engineers faced difficult design challenges in modifying UNIX for effective use on an Intel processing platform. The principal design constraint centered around the limited processing power the Intel chip possessed in the early 1980's. The Intel chip (designed as it was for personal computers) was not nearly as powerful as the enterprise chips used by IBM, Sun, SGI and others in their respective UNIX offerings.
They admit that the early 16-bit Intel chips which Microsoft comissioned Xenix for more than 20 years ago were of "limited processing power". The Xenix system tailored to them was also, unsurprisingly, of "limited processing power".
Only the high price of memory kept Intel chips competitive for mainstream PC users in the 1980s. Since a single megabyte of memory cost over $5000 when the IBM PC was introduced (see the premier issue of PC Magazine, February 1982); a few hundred dollars for the processor was irrelevant at that point. A 32-bit processor's ability to deal with much larger amounts of memory could not be brought to bear until the price of memory came down (by which point Intel had added 32-bit processors to its line and most Unix vendors had begun experimenting with RISC processor designs). However, followers of Moore's Law knew that 32-bit capability would be necessary in future (due to memory prices dropping 50% every 18 months), which is why serious Unix vendors didn't want to waste time on 16-bit platforms. In the early 1980s, Moore's Law was actually an argument _against_ 16-bit Intel processors.
37. Based on the early design constraint of Intel's limited processing power, SCO found an appropriate enterprise market niche for the early versions of SCO OpenServer—single-purpose applications such as point-of-sale control, inventory control and transactions processing, with the highest possible reliability. Intel processors were fully capable of performing these relatively simple, repetitive tasks, and could do so at a lower cost and as reliably as the more powerful enterprise processing platforms sold by the other UNIX vendors, such as Sun and IBM.
The PC's default operating system, DOS, could run any of the "single purpose applications" they list, and it doesn't even multitask. Xenix's selling point was "lower cost" than other Unix systems, especially the ability to squeeze into a tiny amount of expensive memory. SCO seems to think "enterprise" is synonymous with "cheap, low-end market niche". They're proud their OS was capable of performing simple, repetitive tasks at a lower cost than Sun and IBM. To them, this means enterprise.
38. One example of a customer well suited to the earlier version of SCO OpenServer software is McDonald's Corp. McDonald's has thousands of stores worldwide and needs all stores to operate on an integrated computing platform for ease of use, immediate access to information and uniformity. However, the actual computing requirements for each individual McDonald's location are functionally simple—sales need to be tracked and recorded, and inventory functions need to be linked to sales. SCO OpenServer reliably fulfills McDonald's computing requirements at reduced cost.
Again, SCO considers running McDonald's cash registers to be "an appropriate enterprise market niche". This is a task which any modern laptop computer could run in the background without the user at the keyboard noticing. A battery-powered portable CD player has enough computing ability to run a McDonald's cash register. Either they're intentionally lying, or they honestly don't know what "enterprise" means in the operating systems market.
Numerous DOS-based point of sale systems are still for sale, such as apos, posey, Invoice Writer... SCO is basically claming that Xenix was competitive with 16-bit single tasking DOS. Repeating the word "enterprise" in this context renders it meaningless.
39. SCO's business model for SCO OpenServer provides enterprise customers the reliability, extensibility (ease of adding or changing functionality), scalability (ease of adding processors or servers to increase processing power) and security of UNIX—but on inexpensive Intel processor chips. This combination allowed customers to perform an extremely high number of transactions and, at the same time, gather and present the information from those transactions in an economical and useful way for enterprise decision makers.
SCO was not alone in the PC market. In the late 1980s and early 1990s, a number of other companies provided Unix versions and Unix clones for x86 PC systems, and any of them could provide the benefits SCO touts of putting Unix on PC hardware. A decade ago, there was already a fairly diverse PC Unix market.
But by the mid 90s market consolidation in the PC Unix world was already well underway. The Usenet discussion following the February 1995 "going out of business" announcement of one such PC Unix clone vendor, the Mark Williams Company behind the product “Coherent”, is a good sampling of what the PC Unix community understood eight years ago. The users discussing the issue knew exactly why Coherent went under. Already, Linux was a worthy substitute for traditional PC Unix systems, and the Unix clone market was seen as shrinking due to it. The rationale behind open-source development was already fairly mature, as expressed in this post. There was plenty of warning that the proprietary Unix market was in the process of being commoditized by open-source software. SCO had every chance to see its fate coming, and SCO/Tarantella eventually did, which is why it sold its operating systems division to SCO/Caldera.
The history of AT&T's archaic Unix has, for over a decade now, been an example of the “bigger fool” theory. Buy it, then find a bigger fool to buy it from you. AT&T got rid of it because they could not make money from it. Novell chose to exit the Unix market in 1995. SCO chose to exit the Unix market in 2000. Caldera's current lawsuit can be seen as its own form of exit strategy: they want to be purchased, and are making as much noise as they can until somebody buys them to shut them up.
40. The simplicity and power of this “UNIX on Intel” business model helped SCO grow rapidly. SCO gained other large enterprise customers such as CitiGroup, K-Mart, Cendant, Target Stores, Texas Instruments, Walgreens, Merck, Sherwin Williams, Radio Shack, Auto Zone, British Petroleum, Papa John's Pizza, Costco and many others.
When PC hardware became cheap and ubiquitous, SCO/Tarantella failed to capture even 1% market share on it; over 90% of all PC systems shipped with DOS or Windows, leaving Xenix lumped in with also-rans like OS/2 and CP/M-86.
Once again, SCO tries to confuse the meaning of "enterprise computing". SCO has sold cheap low-end systems to large enterprises, but the use they're putting SCO's products to does not qualify as "enterprise computing". Selling napkins to a fancy restaurant does not make one an expert on French cuisine.
41. As Intel's prominence grew in the enterprise computing market, SCO's early version of OpenServer also grew into the operating system of choice for enterprise customers who wanted an Intel-based computing solution for a high volume of repetitive, simple computing transactions.
With no timeframe here, it's hard to tell what they're talking about. With the introduction of the 32-bit 80386 (I.E. 386) processor in late 1985, the x86 line stopped being purely a 16-bit toy, but until IBM first used the 386 in a PC the existence of the processor had little impact on the PC market.[7] IBM announced the availablility of its first 386 processor in the PS/2 line, in April 1987, at the same time it announced AIX for the PS/2, which was a 386 product. SCO shipped Xenix 386 in 1987, and Sun also produced a 386 version of its BSD based SunOS, (before Solaris X86).
42. SCO OpenServer is based on the original UNIX Software Code developed by AT&T, but was modified by SCO for the functionality described above. Thus, while performing single-function applications, SCO OpenServer did so, and continues to do so, with the 99.999% reliability of UNIX.
Translation: Xenix (which became OpenServer) was based on an early version of AT&T's Unix, pared down for early 16-bit IBM PC systems at the behest of Microsoft. It worked about as well as DOS did on the same hardware for running "single-function applications".
43. Over 4,000 separate applications have been written by developers around the world specifically for SCO OpenServer. Most of these applications are vertical applications for targeted functions, such as point-of-sale control for specific industries, inventory control for specific industries, and related functions.
These numbers are unimpressive. Even IBM's OS/2 (a niche operating system IBM gave up on years ago) had more than 4,000 separate applications; DOS and Windows are an order of magnitude beyond that. Linux currently has twice that many open-source projects listed as "Production/Stable" or above in the Trove software map alone.
More to the point, consider SCO/Tarantella's own behavior on the application front before Caldera even bought them. The 86Open Project was an effort to define a common binary executable format to allow the same programs to run without modification on all the various x86 Unix variants. The first meeting of the group took place at SCO's corporate headquarters, on August 22, 1997. Shortly thereafter, a SCO developer named Michael Davidson wrote the utility lxrun, which allowed Linux utilities to run on OpenServer and UnixWare. Davidson made lxrun an open-source project, which allowed Sun to port lxrun to Solaris, and on July 25, 1999 the 86Open group declared victory and dissolved, with the Linux executable format now being the lingua franca of the x86 Unix world.
SCO/Caldera continues to offer lxrun to its customers, along with many hundreds of megabytes of other open-source code, on the SCO Skunkware page. Note that SCO Skunkware was a project of SCO/Tarantella, and predates the Monterey project by a number of years.
As far as SCO/Caldera is concerned, lxrun has been obsoleted by a new product, Linux Kernel Personality for UnixWare, a more comprehensive way to run Linux applications under UnixWare (and presumably a violation of the Linux trademark if they didn't include any Linux code, and a violation of the GPL if they did). It comes built into Open Unix 8.
The Linux Kernel Personality FAQ states:
The LKP feature in Open UNIX 8 represents a significant milestone in SCO's commitment to unify UNIX with Linux for business. Now, the huge number of Linux applications can be deployed on Open UNIX 8. Integrators and users do not have to compromise business requirements because the application they need is not supported on the platform of their choice.
The work to merge UnixWare and Linux predates Caldera's acquisition of SCO; the original Santa Cruz Operation was responsible for the headline SCO revamps UnixWare with Linux features in February 1999. That article contains some interesting quotes about Linux from SCO executives:
"So far as we've seen it's actually helped us," said Greg Schwarzer, director of small and medium business marketing at SCO. "Linux has got the word out that Unix on Intel is a viable alternative to Microsoft."...
Linux, meanwhile, has reinvigorated the Unix market, he said. "It has given a fresh, revitalized look to what people could do in the Intel space. Linux is a Unix movement. Revenues are going up strongly," Schwarzer said.
Adding support for Linux was a pragmatic choice. "We wanted our users to be able to take advantage of a lot of those applications being written for Linux," he said.
Paragraphs 44 through 47 talk about shared libraries, but do not seem to allege that these libraries are either distributed with Linux systems or commonly pirated; nor is there any claim that IBM has misappropriated them. Their main complaint seems to be that nobody wants to buy their SCO emulation module, ergo it must be widely pirated.
44. Much of the functionality of an operating system is made available to application developers by means of “libraries” of code that are supplied by the operating system vendor. These libraries contain many “functions” or “routines” which can be used by application developers to perform various common tasks such as reading or writing a file or opening a new window on the screen.
45. SCO OpenServer, as with many other operating systems such as Microsoft Windows, makes use of a special kind of library called a “shared library.” The code for all of the routines in a particular shared library is stored in a separate file, and this code is loaded into memory “on demand” when an application needs to make use of it. There are several benefits that come from using “shared libraries” - applications can be smaller and use less memory because a single copy of the library code is “shared” by all of the applications that make use of it, and system vendors can easily update the library code in order to fix problems or provide enhanced functionality. A side effect of this is that it is also very easy to make a copy of a shared library.
46. In creating the thousands of SCO OpenServer Applications, each designed for a specialized function in a vertical industry, software developers wrote software code which specifically made use of the SCO OpenServer shared libraries (hereinafter the “SCO OpenServer Shared Libraries”), and thus the presence of the SCO OpenServer Shared Libraries on a particular system is required in order for these applications to be able to run and function correctly.
Without application programs that give an operating system something to do, there's not much point to having most computers even switched on. Third parties wrote a lot of valuable applications, and years ago before Linux rose to prominence and back when the status of BSD was still legally uncertain, they deployed a lot of them on systems like Xenix (which became SCO OpenServer).
When a third party's source code was compiled to create binary executable files to run on a SCO platform, those binaries would try to load SCO's shared libraries to access standard operating system functions. When compiled on another platform (like BSD or Linux), those binaries would instead expect to link up with the BSD or Linux shared libraries. (Versions could even be compiled for other operating systems that had a Unix compatability layer, such as the EMX package for OS/2 or the Cygwin package for Windows. In this case, the resulting binary file would depend on the shared libraries of the appropriate Unix emulation package.) This is an incidental dependency, not inherent in the third party software but simply an artifact of creating binary executable files intended to run on a specific platform.
47. Linux offers a “SCO emulation module,” originally called “iBCS” and now known as “linux-abi” which enables applications which were originally developed to run on SCO OpenServer to be run on Linux. However, in order for these applications to function, the SCO OpenServer Shared Libraries must also be copied onto the Linux system. 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. SCO does not give permission for copying of the Shared Libraries for use outside OpenServer without payment of separate licensing fees.
IBCS stands for the "Intel Binary Compatibility Standard". It was yet another attempt (like 86open and other efforts) to unify the fragmented and incompatible Unix-on-x86 market. Intel pushed the standard in the early 1990s and it's been dead for so long that most references to it have fallen off the web. You can still find a FAQ about it, and an article from 1999 described it as follows:
During the late 1980s/early 1990s Intel promoted a standard binary format for UNIX executables, known as iBCS2 (Intel Binary Compatubility Standard 2). Linux has a kernel module available which provides support for all the various derivatives of iBCS2, so that binaries compiled for a whole range of older UNIXes can run: these include i386 BSD (386BSD, FreeBSD, NetBSD, BSDI/386), SVR4 (Interactive, Unixware, USL, Dell.), generic SVR3, SCO 3.2.x (COFF), SCO OpenServer 5 COFF and ELF binaries, Wyse V/386 (SVR3 with extensions), Xenix V/386, and Xenix V/286. This means that the odds are if you have an existing PC-based UNIX system running some important software (say, an arcane stock management system that isn't supported any more) you may well be able to get it to run under Linux.
Because SCO/Tarantella supported iBCS in UnixWare, it was sometimes possible to run UnixWare applications under other Intel Unix versions, such as Linux, using an iBCS support package for the platform. This would be the only reason to install SCO shared libraries on a Linux platform ala paragraph 46, and in some cases that still wasn't necessary. Linux did have its own shared libraries, and since SCO was presumably providing a standard POSIX/SUS Unix API, in theory adapting them shouldn't be too difficult. In practice, the debugging work required generally wasn't worth the effort, since these deployments were normally aids to migration (to be used only until a native Linux version of the application became available) rather than a permanent solution.
The reason iBCS for Linux has fallen out of use is that it was mainly used to run a small number of proprietary applications that have since released native Linux versions, such as Netscape (which elevated Linux to Tier 1 platform status around the time it, announced its intention to release source code in 1998), and SAS (which ported its software to Linux in 2000). The main iBCS application that Linux users wanted to run, by a wide margin, was Oracle's database product, and when Oracle announced support for Linux in July 1998, iBCS rapidly became irrelevant. In the past 5 years, far more new proprietary packages have been released for Linux than for both of SCO's proprietary Unix platforms combined.
Oracle's announcement of Linux support came the same month Linux creator Linus Torvalds was on the cover of Forbes. As the leading database vendor, Oracle's announcement prompted IBM to announce a Linux version of its own database, DB2, two months later. That announcement was IBM's first official acknowledgement of or support for Linux at the corporate level[8], and it was a reactive rather than proactive measure.
In 2003, SCO no longer has interesting applications that Linux lacks. Since at least 1998, before IBM first became even peripherally involved with Linux, major software development houses have shipped Linux versions of their software, often instead of SCO versions. SCO has responded by creating technologies like lxrun, and the Linux Kernel Personality for UnixWare, to run Linux applications on SCO's operating system. SCO/Tarantella realised this was a losing strategy and got out of the proprietary Unix business in 2000.
Recently, SCO resurrected the issue of iBCS support as part of its "System V for Linux" product, the first deliverable of the SCOsource initiative, which is the newly launched division of SCO that is driving the lawsuit against IBM. The presentation describes System V for Linux as "SCO's Shared UNIX Libraries from OpenServer and UnixWare for use with Linux", and claims that two answers to the question "Why License SCO's Intellectual Property" are "Strengthen Linux by licensing value-add IP" and "Increase UNIX application use on Linux". It also claims:
In the late 1980s, SCO created an open specification called ibcs2 (Intel Binary Compatibility Standard) to allow UNIX applications to run on Intel standard hardware using SCO's shared libraries.
Because ibcs2 is an open specification, the Linux community was able to freely copy this and rename it Linux ABI or Linux Application Binary Interface.
In order to run UNIX apps on Linux, customers must have Linux ABI and SCO's shared libraries.
SCO's shared libraries could never be licensed outside of the SCO operating systems until today.
The third paragraph of that is wrong: in order to run UnixWare applications under Linux, SCO's shared libraries might be necessary. When attempting to run applications for other x86 versions of Unix, such as the x86 versions of AIX or Solaris, SCO's shared libraries would be of no use whatsoever.
It's interesting to note that they claim to have invented the Intel Binary Compatibility Standard, rather than Intel. Also, the most recent document they can find explaining how to run a SCO binary on Linux explains how to run WordPerfect 5, from 1995. Caldera paid to have WordPerfect 6.0 ported to Linux, which is why subsequent versions have also been available for Linux. Caldera's presentation also references the retention of Boies, Schiller and Flexner by SCOsource (the lawfirm retained to persue the current suit against IBM) while at the same time it hawks a Linux product. Apparently, this attempt to profit from Linux and the IBM lawsuit were launched more or less simultaneously.
SCO points out that System V for Linux is the first time they've licensed their shared libraries seperately from their operating system, but skips over the fact that people who legally owned a copy of a SCO operating system but chose to migrate to Linux could legally use those shared libraries under Linux, because they do have a license to own them and copyright law cannot pevent the re-use under the “first sale” doctrine. (It could also be considered an example of "space shifting".) The only users likely to have a significant number of legacy SCO applications they'd want to run on a new platform are users who migrated from a SCO platform to Linux, and thus have a SCO license they are otherwise no longer using.
48. While the original SCO OpenServer operating system performs with all the reliability and dependability of other UNIX systems, it was originally designed for the initially low processing power of Intel chips. Therefore, SCO OpenServer does not offer the same level of multiprocessor capabilities that other versions of UNIX offer.
So why did they spend so many paragraphs insisting it's an "enterprise operating system"?
49. During or about 1992, SCO's predecessor in interest, Novell, Inc. (“Novell”), acquired from AT&T all right, title and interest in and to the UNIX software code, the AT&T Software and Sublicensing Agreements, the copyrights and related and ancillary products for $750 million in Novell stock. For branding purposes, Novell renamed UNIX as “UnixWare.”
Novell acquired Unix System Labs (USL), which AT&T had spun off as an independent company. Novell bought out AT&T's stake in USL. Novell didn't acquire "all right, title and interest" to Unix from AT&T because AT&T didn't have it to give: a large number of existing licenses had been granted to companies like IBM delegating rights to create derived works that belonged to the licensees, not to AT&T. For years, far more Unix development had taken place outside of AT&T or USL than within it.
This period in Unix history is still remembered for fragmentation, divergence, and increasing incompatibility, because licensees were not sharing new developments with each other. One reaction against this was standardization efforts like Posix, OSF/1, the X/Open portability guides, the Single Unix Specification, and The System V Interface Definition. Most modern unix-like systems, including Linux, are compliant with many or all of these published standards. Technically speaking, even Microsoft's Windows NT has achieved nominal compliance with some of them.
The largest development branch AT&T didn't own was from the University of California at Berkeley. The BSD lawsuit was settled on Novell's watch, and settled in Berkeley's favor, not only because Berkeley exercised their right to create derived works until they had replaced the last of the original System V code, but also because AT&T and USL had contaminated System V with code they didn't have rights to. Derived code belonging to licensees wound up in System V, and it came out in court. Novell ordered USL to settle the lawsuit, and unloaded System V shortly afterwards.
50. On or about September 19, 1995 the Santa Cruz Operation acquired all right, title and interest in and to UNIX and UnixWare source code, the AT&T Software and Sublicensing Agreements, the copyrights, claims arising after the closing date against any party and all related and ancillary products and rights from Novell, excepting only the right to certain existing ongoing royalty payments which was retained by Novell.
SCO did not get the Unix trademark, which Novell gave to The Open Group, and Novell has disputed the copyright assignment (a question which simply had not arisen for seven years between the transaction and this lawsuit). SCO also continued to pay a significant amount of royalties to Novell, $19 million in the first quarter alone. Novell's 1996 first quarter 10-Q records the transaction as follows:
In December 1995, Novell sold its UNIX and UnixWare product line to the Santa Cruz Operation, Inc. (SCO). The Company realized a small gain and recorded $19 million of royalty revenue from this transaction in the first quarter of fiscal 1996. Under the agreement, Novell received approximately 6 million shares of SCO common stock, resulting in an ownership position of approximately 17% of the outstanding SCO common stock. The agreement also calls for Novell to receive a revenue stream from SCO based on revenue performance of the purchased UnixWare product line. This revenue stream is not to exceed $84 million net present value, and will end by the year 2002. In addition, Novell will continue to receive revenue from existing licenses for older versions of UNIX System source code.
The "older versions", on which SCO still pays royalties to Novell, include existing System V licensees. SCO/Caldera's most recent annual report reports a royalty payable to Novell of $1.4 million for fiscal year 2002. If SCO has free and clear rights to Unix, why are they still paying royalties on it to Novell?
51. From and after September 1995, SCO dedicated significant amounts of funding and a large number of UNIX software engineers, many of whom were original AT&T UNIX software engineers, to upgrade UnixWare for high-performance computing on Intel processors.
52. By approximately 1998, SCO had completed the majority of this task. That is to say, UnixWare had largely been modified, tested and “enterprise hardened” to use Intel-based processors in direct competition against IBM and Power PC chips, the Sun SPARC chip and all other high-performance computing UNIX platforms for all complex computing demands. The term “enterprise hardened” means to assure that a software product is fully capable of performing under the rigorous demands of enterprise use.
Translation: SCO/Tarantella invested three years of work to try to breathe life into the obsolete product Novell had just unloaded on them. They tried to upgrade it to some sort of relevance in the marketplace, and the result was to turn UnixWare 2 into UnixWare 7, which the industry perceived as a non-event. In 1999 The UnixWare "Business Edition" was released, and a quote from an article about it states, "The Business Edition supports a single-CPU system with up 4 gigabytes of memory and up to five users." During this period, top of the line x86 "Xeon" systems from Intel already supported up to 8 processors and 64 gigabytes of memory.
In 1998, just as SCO/Tarantella stopped throwing good money after bad, Linus Torvalds appeared on the cover of Forbes magazine, Oracle announced support for Linux, Netscape released its source code (citing Linux as an inspiration), and Linux-based "Beowulf" clusters showed up in the top 500 supercomputers list and were covered by CNN. All of this occurred before September, when IBM first supported Linux — not by contributing directly to development of the operating system, but by releasing a Linux version of IBM's DB2 database application in response to database market leader Oracle's Linux initiative (which has continued). Linux hit the big time long before IBM noticed it.
SCO/Tarantella had noticed Linux as early as 1995, when Linux put Coherent out of business. In 1998 its responses to Linux included iBCS, lxrun, and the 86open project. It also stepped up its "skunkware" project, to collect and distribute open source for SCO platforms (even shipping skunkware CD's with SCO's operating system). SCO's "Open Source Program Architect", Ronald Joe Record, gave a very interesting slide presentation about SCO's open-source efforts at SCO Forum 98. Major releases of SCO Skunkware were made in 1998, 1996, and earlier.
SCO/Tarantella's co-founding of the Monterey project in October 1998 was a second attempt to make their proprietary Unix technology relevant to enterprise customers. If their own in-house efforts had succeeded, they would not have needed to partner with IBM, Sequent, and others.
The three years of development work in pargraphs 51-52 have no bearing on the historical System V code other vendors had already licensed and upgraded on their own in the 1980s and early 90s. Any development work done by SCO in 1995 would merely create yet another derivative work, in this case UnixWare 7.
53. SCO was ready to offer large enterprise customers a high-end UNIX computing platform based on inexpensive Intel processors. Given the rapid growth of Intel's performance capabilities and Intel's popularity in the marketplace, SCO found itself in a highly desirable market position. In addition, SCO still had its SCO OpenServer business for retail and inventory-targeted functions, with its 4,000 applications in support.
54. Prior to the events complained of in this action, SCO was the undisputed global leader in the design and distribution of UNIX-based operating systems on Intel-based processing platforms.
This statement is simply false. See the original OSI position paper for a detailed refutation. By the end of 1998 SCO had added support to its products for the Intel Binary Compatibility Standard (iCBS), participated in the 86open project, and created lxrun. None of these actions would have made sense for an “undisputed leader”.
55. As SCO was poised and ready to expand its market and market share for UnixWare targeted to high-performance enterprise customers, IBM approached SCO to jointly develop a 64-bit UNIX-based operating system for a new 64-bit Intel platform. This joint development effort was widely known as Project Monterey.
The translation of "poised and ready to expand" is something like "scratching its head wondering why nothing had happened". SCO/Tarantella had never previously been in the same league as Sun, HP, and IBM, and its development efforts did not suddenly elevate it there. Linux was already eating into proprietary Unix vendors' marketshare on 32-bit x86 machines, prompting Sun to give away free licenses to Solaris for x86. And Linux itself was not the only reason for the erosion of SCO's market; the success of Linux and FreeBSD helped pave the way for Apple's may 1998 announcement of a new OS development strategy, MacOS X based on a combination of FreeBSD and Carnegie Mellon's Mach microkernel. (Here's an article on the convoluted history of MacOS X). Apple's move to Unix also cut into the low and midrange Unix markets, SCO's bread and butter. Their attempted move to high end "enterprise computing" was an exit strategy from their rapidly deteriorating existing market.
SCO/Caldera seems to be attempting to imply that project Monterey was some kind of bait distracting them at a crucial moment, and that if it hadn't been for IBM, UnixWare would have blossomed into a success. The fact that UnixWare continued to be developed and marketed during Monterey development and continued to be a dud is a counter-argument. SCO/Caldera joined Monterey because their own in-house development efforts had not achieved significant enterprise market share. They wanted IBM's help. They jumped at the chance to get it.
56. Prior to this time, IBM had not developed any expertise to run UNIX on an Intel processor and instead was confined to its Power PC processor.
Paragraph 56 is completely wrong. AIX for PS/2 was announced at the PS/2 launch in 1987. IBM also shipped AOS, a BSD derivative, for PS/2 systems. IBM's first Unix product for its x86 PC shipped in 1983 (as described in the response to Paragraph 26).
In addition, the statement is intentionally misleading. Although Itanium was from Intel, it was a very different design from Intel's traditional x86 processor line. Monterey was an effort to port Unix to a completely new hardware architecture, a 64-bit processor with a new instruction set (based on VLIW design principles which, like RISC, IBM had invented years before). Itanium (originally code-named Merced) was jointly designed by Intel and HP. Writing code for the 64-bit Itanium processor has more similarity with writing code for 64-bit RISC Power PC processors than writing for 32-bit CISC x86 processors.
Monterey was not the first such 64-bit processor either. The first 64-bit RISC processor, the DEC Alpha (now owned by HP due to its purchase of Compaq, which had previously purchased DEC) was introduced in February 1992. The first 64 bit Unix for this processor, DEC OSF/1 (later renamed Tru64 Unix) was introduced in March 1993. It had nothing to do with SCO.
The article 27 years of IBM RISC points out that IBM's own internal development led to the release of a 64-bit PowerPC processor and a 64-bit version of AIX (4.3), in 1997. The article says:
The RS64 (also known as Apache) is the first 64-bit PowerPC RISC processor (October 1997). The RS64 is a superscalar processor optimized for commercial workloads... IBM brings 64-bit technology to the market introducing the RS/6000 Enterprise Server model 7017-S70 (125 MHz, code named Raven), the first 12-way SMP system, and AIX Version 4.3.
IBM had more experience with 64-bit computing than SCO did, having designed and marketed its own 64-bit processor and a version of AIX that could run 64-bit applications.
57. SCO, on the other hand, had over 15 years of expertise in adapting UNIX to Intel based systems. Moreover, SCO had spent the previous 18 months working closely with Intel to adapt its existing UnixWare product to work on the new 64-bit Intel processor. That project, known as "Gemini-64," was well underway when work on Project Monterey was started. In furtherance of, and in reliance on, IBM's commitment to Project Monterey, SCO ceased work on the Gemini-64 Project and expended substantial amounts of money and dedicated a significant portion of SCO's development team to Project Monterey. Specifically, plaintiff and plaintiff's predecessor provided IBM engineers with valuable information and trade secrets with respect to architecture, schematics, and design of UnixWare and the UNIX source code for both 32- and 64-bit Intel-based processors.
Over 15 years ago SCO was adapting Xenix to the 16-bit 8086 and 286 processors in the original IBM PC and AT lines, which were less powerful than a modern cell phone. This has no relevance whatsoever to modern "enterprise" systems. The first 32 bit processor in the x86 line, the 386, supported over a dozen Unix ports, including AIX for the PS/2 from IBM.
SCO is trying to confuse "Gemini 64" with the three years of development they mentioned in pargraphs 51-52. The purpose of the development in pargraphs 51-52 was to turn UnixWare 2, which SCO/Tarantella purchased from Novell in 1995, into UnixWare 7, released in 1998. This development resulted in a product release for 32-bit x86 Intel processors (the 386, 486, and Pentium lines).
The Gemini-64 project was an experimental effort at SCO to play around with Intel's unfinished 64-bit design (code named "Merced", and released as "Itanium"), which could only be tested on emulators since Intel hadn't manufactured any actual 64-bit processor chips yet. SCO eventually admitted they were in over their heads, canned Gemini-64, and sent the Gemini engineers off to work with IBM.
By the way, does SCO really claim to have trade secrets related to the architecture and schematics of Intel's new hardware design? (Wouldn't these be Intel's trade secrets, which IBM could get from Intel? If IBM could NOT get them directly from Intel, wouldn't SCO be violating its agreements with Intel in passing them on?) The most charitable reading is that SCO is trying to inflate its importance, or simply confusing two issues again.
58. By about May 2001, all technical aspects of Project Monterey had been substantially completed. The only remaining tasks of Project Monterey involved marketing and branding tasks to be performed substantially by IBM.
Except that market interest in Monterey never materialized, in large part because the Itanium hardware was a flop, completely unsuccessful in the mainstream computer market. To read the gory details of the Itanium disaster, see Appendix D at the end of this document.
Beyond that, the initial point of the Monterey project was another standardization effort, like the Open Software Foundation of the late 1980s. Periodically, when the proprietary Unix market gets too fragmented, a group of vendors band together for a while, and after a while collapse into bickering and pointless infighting. The Open Software Foundation was originally a joint effort between leading Unix vendors including IBM, Hewlett Packard, and Digital Equipment Corporation (now part of HP, via Compaq), in reaction to AT&T's purchase of an equity stake in Unix vendor Sun Microsystems. (Everybody but Sun and AT&T formed an opposing camp.) OSF eventually collapsed, and their operating systems release (OSF/1) became Tru64 Unix, a product for the DEC Alpha processor. Other theoretically similar standardization efforts predating Monterey include Unix International, and the groups behind Posix and the Single Unix Specification. (The second edition of the book "Life with Unix" talks about all this in detail.)
Monterey was another such standardization effort, as explained in Appendix D.
59. On or about May 2001, IBM notified plaintiff that it refused to proceed with Project Monterey, and that IBM considered Project Monterey to be “dead.”
For entirely justifiable business reasons. The Itanium was a disaster, IBM had purchased Sequent for its NUMA technology, and Linux had gone mainstream around 1998. Project Monterey was an obvious failure long before May 2001.
Three years ago, the lead author of this commentary, Rob Landley, summarized the Monterey situation in a comment attached to a July 19, 2000 article, which was one of many articles pointing out that the reason the Linux distributor Caldera purchased SCO was not to get its technology, but to get an international distribution channel through which to sell Linux. (The fact that Caldera immediately renamed itself "Caldera International" might have been another hint about what part of the deal excited them.)
60. AT&T Technologies originally licensed the UNIX operating system software code to hundreds of software licensees, including defendant IBM, for the UNIX operating system software source code, object code and related schematics, documentation and derivative works (collectively, the “UNIX Source Code”). To protect the confidential and proprietary source code information, these license agreements, as detailed below, contained strict limitations on use and distribution of UNIX source and binary code.
SCO/Tarantella contributed code to Linux in 1999, and was thinking of putting out its own Linux distribution until Caldera purchased its operating systems division. Some news reports state that SCO/Tarantella's moves into Linux were one reason for the Caldera acquisition.
SCO/Caldera was a Linux distributor, and repeatedly announced plans to put Unix technology into Linux. What if one of Caldera's Linux engineers believed them? Caldera employed a number of high-profile Linux engineers, such as Christoph Hellwig, who was in the top ten list of contributors to Linux 2.4 while working for Caldera in Germany.
Christoph Hellwig is of the opinion that SCO probably didn't intentionally put UnixWare code into the Linux kernel:
"Especially given that the kernel internals are so different that you'd need a big glue layer to actually make it work and you can guess how that would be ripped apart in a usual lkml [Linux Kernel Mailing List] review :)"
I.E. he believes that SCO/Caldera had "better things to do than trying to retrofit UnixWare code into the linux kenrel[sic]" because if SCO/Caldera had submitted UnixWare code to the Linux kernel, it would have been rejected on technical grounds. (This is further support for the contention that UnixWare is obsolete crap.)
Christoph remains very active in Linux kernel development since having left SCO/Caldera, and is only one of thousands of people outside of IBM who have legally worked with traditional proprietary Unix code, but have left proprietary Unix behind in favor of Linux. SCO claims that code licensed to thousands of entities by many different companies over a period of decades has never legitimately leaked once. The co-author of this document, Eric Raymond, doesn't think so: his No Secrets page has already collected over 100 reports from old-time Unix administrators who still have copies of System V-derived Unix source code in their possession.
Hundreds of thousands of developers have had legitimate access to this code for two decades, and SCO claims that no employment agreement ever missed telling them to destroy their copies when they left the company, no company ever went bankrupt owning Unix code and had its unformatted hard drives liquidated at auction, no disks or printouts were ever thrown out without being shredded and fished out of the trash...
AT&T sold educational licenses for Unix source code to universities at a considerable discount over commercial use, yet how many professors made their students sign a non-disclosure agreement? How can something explicitly licensed for educational purposes and knowingly taught to all comers in a university classroom be considered a "secret", by any definition?
And with all these thousands of different organizations that had access to this code, why is SCO picking on IBM specifically? (Other than IBM's deep pockets?)
61. When SCO acquired the UNIX assets from Novell in 1995, it acquired all right, title and interest in and to the UNIX operating system technology, including all claims against any parties relating to any right, property or asset used in the business of developing UNIX and UnixWare. As a result of this acquisition, SCO became the authorized successor in interest to the original position of AT&T Technologies with respect to all licensed UNIX software products.
Just like the DR-DOS lawsuit in 2000. None of Caldera's predecessors felt like suing, but Caldera buys up dead technologies to see if it can squeeze a lawsuit out of them. Microsoft was motivated to settle by the DOJ antitrust loss. IBM does not have a history of misappropriating code. (And Unix is one of thousands of different technologies they've licensed from third part