Our previous articles on day 2 covered the second day of Chris Sontag's testimony (which I put together with his first day, for logic and convenience), then Darl McBride's, and then Greg Jones' testimony. Those were Novell's witnesses, the ones it called.

Novell's case having been fully presented, it was SCO's turn to put on its case, each side trying to present evidence to support their opening statements. The first SCO witness called is John Maciaszek. We've seen him before on paper, and so has Judge Dale Kimball. But this is our first opportunity to watch him in person.

I think by the end of his testimony, SCO's argument that UNIX System V and UnixWare are the same thing completely unravels.

Maciaszek was a director "who has worked for USL, Novell, Santa Cruz and SCO as a Contract Manager in the UNIX licensing groups". He was director of operating system product management at SCO . But throughout his career, he was product manager for UnixWare, as you can verify by checking his testimony, introduced into the SCO v. Novell litigation as Exhibit 34 [PDF] attached to a Declaration of Edward Normand in April of 2007. He testifies here that he's semi-retired, but SCO told the bankruptcy court [PDF] earlier that he was terminated on January 31 and paid severance.

It's by the end of his testimony that I think SCO's story about Unix System V and UnixWare being one and the same gets revealed as bogus. You'll recall that Darl told the court with a straight face that UnixWare is the trunk of the tree of Unix, nay not just the trunk but the branches and the leaves, and he added that there is only one way to get UNIX code -- you had to buy UnixWare:

And so if somebody came to the company and said, we want to get the core intellectual property to UNIX, and we want to take a license for that, for example, IBM did that with us in 1998, we said, okay, if you want to get core access to the UNIX intellectual property or the trunk code, the way you do that is through a UnixWare license.

I'll talk about some of what he says about IBM in 1998 in more detail in a minute, but on the Unix/Unixware part, here's what Maciaszek is asked and answers:

Q. How much code of UnixWare would there need be to trigger the UnixWare royalty as opposed to one of the older royalties?

A. Well, that's the concept. The one-line-of-code concept said if you licensed UnixWare 2.1 and it was the latest release, if you took a single line of code from UnixWare and included it in a derivative work, your derivative work as you distribute it would be subject to the terms, conditions and obviously royalties which were part thereof of the UnixWare 2.1 schedule. So you'd pay based on UnixWare if you took one line of code from UnixWare sources and put it in derivative work.

Q. That's why it's called one line of code?

A. That's correct.

Q. So even if 99 percent of the code in that derivative product was an older System V version, if it was one line of UnixWare, it would trigger UnixWare royalty?

A. That is correct.

Q. And under the APA, are you aware whether any UnixWare royalties had to be remitted to Novell?

A. No UnixWare royalties remitted to Novell.

Excuse my simplicity, but as I read it, is he admitting that SCO found a tricky way to avoid paying Novell what the APA contemplated for SVRX? Under a one-line-of-code system, and one which made it impossible to get the code unless you bought UnixWare, exactly when *would* Novell get paid under the APA? See what I mean?

Yet, as Judge Kimball ruled in August, those royalties were contemplated as a significant part of the purchase price. And we know it has been paid something over the years, so it's hard to parse if Unix System V and UnixWare are the same thing and particularly if there is no way to get System V code without obtaining UnixWare. It's circular. It puts Novell in a Catch-22.

And at exactly what point in time did it become impossible to get System V code without UnixWare? IBM got it without UnixWare in the beginning. So did all the hardware guys. So, was it post-APA that they pulled the switch? If so, would that be a violation of the contract? Or is it all dancing baloney, this new story, as Novell argues to the court in its closing arguments?

When SCO first sued IBM for beellions, they told the world that System V code was holy and precious, and SCOfolks imagined buying second homes, castles in Spain, so to speak, from all the coin they thought they'd make from it. Now, after the court ruled those copyrights are Novell's, the SCO story is that it's worthless, you can't buy it, no one wants to anyway, it's really UnixWare after all, etc.

By the way, if they are the same thing, why did SCO in 2003 copyright Unixware 7.1.3 and UNIX System V separately? You tell me. But they did, as you can see on our Contracts page. So did Novell. That would indicate to my mind that there is code in each that is not identical. No? I think the seesaw is so tilted on Novell's side that SCO's feet are dangling in the air. But you can reach your own conclusions.

Now, SCO might argue that the SVRX royalties contemplated were only for prior purchases or only for binary something or other -- remember all those impenetrable arguments before Judge Kimball's August 10th Order? They lost that argument, and my memory was they were not supposed to argue it again at trial. But with Maciaszek's testimony, I suspect that is what they did. I remember that SCO itself quoted the following words from the APA in its response to IBM's motion for partial summary judgment:

The APA defined the "Business" as "the business of developing a line of software products currently known as Unix and UnixWare, the sale of binary and source code licenses to various versions of Unix and UnixWare, the support of such products and the sale of other products which are directly related to Unix and UnixWare."

Emphasis added by me, to highlight that the APA contemplated being able to buy either or both, separately. In fact, in the January 23, 2007 hearing in SCO v. Novell, on Novell's motion about getting the money from the Microsoft and Sun deals, we learned that Microsoft licensed both UNIX SVRX and UnixWare, and we learned it from the lips of SCO attorney Stuart Singer. How could that be, if SCO's story is true that the only way to license SVRX is to license UnixWare instead? And if it were, when did that change, and was SCO allowed to change its business unilaterally, in light of its fiduciary responsibilities?

Getting back to Darl's testimony, first, when IBM and Santa Cruz got together in 1998, IBM already had a license to Unix System V. They had built AIX already by the end of the 1980s. The 1998 deal was a joint project, where both parties and a few others would each contribute some code to the project, Project Monterey. SCO contributed UnixWare 7 to the project. Here's the license with IBM and Santa Cruz, and it's called "JOINT DEVELOPMENT AGREEMENT AGREEMENT NUMBER 4998CR0349". Here's the opening part, the so-called Whereas clauses, where the contract tells you why the parties want to sign the contract:

Whereas, IBM and SCO believe that each company has complementary skills, experience and technology to extend and evolve existing UNIX operating systems designed to operate on the 32-bit and 64-bit Intel architecture platforms;

Whereas, pursuant to this Agreement, IBM and SCO desire to undertake specific development projects to design and further develop an IA-32 Product and IA-64 Product as further defined below. Through these projects the parties expect to enable innovative new open systems computer technologies and products more rapidly and efficiently than either party could achieve independently;

Whereas, each party expects to market products based on this development work and to work toward the goal of creating and participating in a high volume UNIX marketplace, bringing significant benefits to each party, its customers and the computer industry;

Now, therefore, in consideration of the mutual provisions contained herein, SCO and IBM (the "Parties") agree to the following terms and conditions.

So IBM didn't need SVRX and it certainly didn't license UnixWare 7 to get it. It had it already and it also had the independent skills and code necessary to extend Unix code, which it had already licensed, as you can see in the Whereas clause on a contract signed by Santa Cruz also. This was a joint project, so they each contributed skills and code to the effort, and they decided SCO's contribution would be UnixWare 7. IBM was bringing their Power PC chip expertise and marketing fu, among other things. It had nothing to do with the SVRX/UnixWare APA issues. In Project Monterey, the idea was to create a joint product. And it was a cross license of two branches from the UNIX System V trunk, AIX and UnixWare. Both parties had access to the trunk already. Here's that bit from the contract:

1.9 "IA-32 Product" shall mean the UNIX operating system that is designed to run on Intel architecture and compatibles and which consists of SCO's UnixWare 7 with the addition of Licensed IBM Materials and any additional Project Work developed under this Agreement.

1.10 "IA-64 Product" shall mean the UNIX operating system that is designed to run on Intel architecture and compatibles and which consists of IBM's AIX operating system with the addition of Licensed SCO Materials and any additional Project Work developed under this Agreement.

Darl's words are, frankly, hard to match to real life, shall we say? I don't know if he misspoke or if he was trying to mislead but for sure his words are misleading if you came away with the idea that IBM approached SCO and went to get a license for UNIX, and in order to get it, had to buy UnixWare. It was a cross-license, not IBM asking for access to Unix. It got that in the '80s. In the 1985 license, whereby IBM got a license for "UNIX* System V, Release 2.0", you can see that this wasn't even the first such agreement, and it wasn't the last, as you can verify on our Contracts page under the IBM heading. Note in the Supplement, if you scroll down to the part that lists what IBM had access to under the agreement, it mentions a 1982 agreement between IBM and AT&T already existed:

All CPUs that are SOURCE CPUs under the prior January 1, 1982 Software Agreement, as Modified, Between Our Companies Relating to UNIX System V, Release 2.0 and other UNIX Operating Systems, for which agreement this agreement has been substituted. All fees for such CPUs have been paid pursuant to such prior agreement.

That's exactly what Novell told the court was the system in its opening statements, remember? You got the earlier code free if you had already paid for it. You might want to remember that detail, as you go through the Maciaszek testimony and weigh its value, because he testifies that earlier versions of Unix were just thrown in free for any new customer. Here you can clearly see, it references that IBM could use the earlier version too because it had already paid for it, which is what Novell's attorneys brought out in their part of the trial.

We've seen Mr. Maciaszek before, and so has the Honorable Dale Kimball, as I said. Here's when. Novell filed evidentiary objections to his testimony, that same Exhibit 34 already mentioned, along with some sealed depositions, and asked that it be thrown out. It did so a second time. The testimony was in support of SCO's opposition to Novell's various summary judgment motions. His testimony, as described by SCO in its motion for reconsideration of the Utah court's August 10th order, involved this aspect of the contract dispute:

Second, the uncontroverted testimony demonstrates that the parties always treated revenues from UnixWare Licenses that included incidental licenses to SVRX as UnixWare revenues under Schedule 1.2(b) of the APA, not as SVRX Royalties....

John Maciaszek has been involved in the UNIX business for almost two decades. (Maciaszek Decl. (12/11/06) ¶ 2.) He testified that Novell's interest in UnixWare Licenses that included an incidental license for prior SVRX products was always treated under the UnixWare Royalty provisions of Schedule 1.2(b), and that interest "expired in 2002." (Id ¶ 27.)

Kimball was not moved, and he did not reconsider his ruling.

Once again, our thanks go out to Groklaw's papafox, who developed the script that makes it possible to do transcripts with line numbers without going crazy. Sometimes in court documents, you need to know the line numbers to be able to follow along. A judge might reference testimony by page and line number, like this, as an example: Plaintiff's Deposition 50:12. You know that the 12 is the line number. We always do transcripts first without line numbers, so those who rely on screen readers don't have to hear numbers read aloud for each line, but I also always try to do the versions with line numbers too, so you can reference them when they are cited. Thanks to papafox, it's so much easier now.

**************************

389

390

391

392

393

394

395

396

397

398

399

400

401

402

403

404

405

406

407

408

409

410

411

412

413

414

415

416

417

418

419

420