Did SCO get Linux-mob justice?September 10, 2007: 6:14 AM ET
Once in awhile a judicial ruling comes down that's so wrong at such a basic level that you're just left scratching your head.
When U.S. District Judge Dale A. Kimball of Salt Lake City threw out most of SCO Group's (SCOX) case against Novell (NOVL) on August 10 -- effectively dooming most of SCO's claims in closely related cases against IBM Corp. (IBM), AutoZone (AZO), and Red Hat (RHT), too -- his 102-page ruling was greeted with widespread rejoicing and I-told-you-so's. (I was headed out of the country for vacation at the time and am only now digging out from under.)
Understandably, few people mourned SCO's humiliating defeat. In a series of incendiary lawsuits and letter demands in 2003, SCO sought licenses from at least 1,500 companies that used or distributed Linux, claiming that, by doing so, they were either breaching UNIX-related contracts or infringing UNIX copyrights, both of which SCO claimed to own. The demands enraged not just the Linux developer community, but many Fortune 500 companies that had become big Linux users and champions. (For a feature story I wrote about the disputes in May 2004, click here.)
Still, as a piece of judicial craftsmanship, Kimball's work falls squarely within that rare category I describe in the first sentence of this post. Here's Kimball's ruling. (Novell and SCO declined to comment for this article, citing the imminence of the trial of the few remaining issues in the case, which starts September 17.)
The problem is not that Judge Kimball's view of the facts is wrong; it might not be. His judgments about which testimony to believe and which not to believe are, in fact, plausible. So are the inferences he draws from that testimony about how he should interpret the monumentally gnarly, self-contradictory, and, in my humble opinion, ambiguous 1995 contract that lies at the heart of the case. If SCO had asked to have its case tried before a judge (a "bench trial"), and if judge Kimball had then held that trial -- so he could see the witnesses testify in the flesh and make informed judgments about their live demeanor -- his ruling would make perfect sense and I'd have no objection to it.
But SCO didn't ask for a bench trial, and Judge Kimball never held one. SCO asked for a jury trial, and Judge Kimball was, therefore, only ruling on Novell's pretrial motion for summary judgment. And as any second-semester law student knows, a judge can grant such a motion only when, as innumerable courts in every state and federal jurisdiction have repeatedly written, "the evidence, viewed in the light most favorable to the party opposing the motion [i.e., SCO, in this situation], shows there are no genuine issues of material fact." (If that weren't the rule, our Seventh Amendment right to a civil jury trial would be a hollow joke.)
In ruling on such a motion, a judge cannot "act as the jury and determine witness credibility, weigh the evidence, or decide upon competing inferences," according to the well-worn case law. You will find these or equivalent boilerplate recitations of the applicable law pasted somewhere into damn near every summary judgment ruling you will ever come across, with one conspicuous exception: Judge Kimball's August 10 ruling in the SCO case. (After the ruling, the only claims left in the case were of a nature that do not entitle a party to a jury, and on Friday, September 7, Kimball granted Novell's request to hold a bench trial on those. But there's no dispute that SCO would have been entitled to a jury on the claims that were tossed out.)
For those who came in late, the case is principally about an "asset purchase agreement" signed in September 1995 in which Novell sold something -- nobody's sure exactly what any more -- to a company called Santa Cruz Operation for $125 million plus certain royalty streams. (Santa Cruz later sold whatever-it-was-that-it-obtained-from-this purchase to a company called Caldera, which later changed its name to SCO.) SCO says Santa Cruz (and, ultimately, SCO) got from Novell the entire UNIX operating system business, including copyrights, while Novell says that Novell actually withheld the UNIX copyrights at the last minute, and only sold Santa Cruz, essentially, a license to take the UNIX code, use it, and make and sell new products out of it.
The then-CEOs of both Santa Cruz and Novell (yes, of Novell too) each supported SCO's position in their testimony -- i.e., the position Judge Kimball rejected without even letting a jury hear it. Each former CEO said that it was his understanding that Novell had sold Santa Cruz the entire UNIX operating system business, including copyrights. Here's how Novell's then CEO Robert Frankenberg testified:
Q. Was your initial intent in the transaction that Novell would transfer copyrights to UNIX and UnixWare technology to Santa Cruz?
Q. Was that your intent at the time when the APA was signed?
Q. Was it your intent when that transaction closed?
Q. And did that remain your intent, as you view it, at all relevant times?
Q. That never changed?
That was also the view of Novell's (yes, Novell's) then chief negotiator, Ed Chatlos. In fact, it appears to have been the understanding of nearly every negotiator on both sides of the table, with two important exceptions.
The exceptions were Novell's then-general counsel David Bradford and Novell's then outside counsel Tor Braham, of Wilson Sonsini Goodrich & Rosati, who each testified that toward the end of the negotiations Bradford told Braham to withhold the Unix copyrights from the deal, either because Santa Cruz couldn't pay enough money or because they feared Santa Cruz might go bankrupt.
This understanding was news to Burt Levine, another member of Novell's inhouse legal team, however, who said he had also reviewed and revised drafts of the contract.
Q. Mr. Levine, from the time of the [asset purchase agreement] in 1995 until you left Santa Cruz in 2000, did you ever hear anyone, whether inside or outside of Santa Cruz or inside or outside of Novell, say that Novell had retained the UNIX or UnixWare copyrights?
Q. If you had heard anyone make such a statement, would that have been a surprise to you?
A. Very much so, yeah.
The Bradford/Braham recollection was also contradicted by a member of Santa Cruz's inhouse legal team at the time, paralegal Kimberlee Madsen, who was closely involved in the meetings and negotiations. "It was always my understanding," Madsen wrote in her declaration, "that the UNIX source code and its copyrights were part of the assets Santa Cruz purchased and were transferred to Santa Cruz at the closing in December 1995. I do not recall anyone in the negotation teams ever saying, or suggesting, that Novell would retain any UNIX copyrights. The negotiation team for Santa Cruz never discussed the possibility, as far as I am aware, that Novell sought to retain any UNIX copyright."
Again, in choosing to believe that Bradford and Braham were more credible, closer to the action, remembered the situation better, or what have you, Judge Kimball drew plausible and defensible inferences -- for a juror. But a judge isn't allowed to do that in ruling on a summary judgment motion.
Likewise, Judge Kimball weighed the credibility of then-Novell CEO Frankenberg's testimony, and found it wanting. He writes that Frankenberg was "self-contradictory" at times (he cites no examples), and that portions of his testimony were contradicted by certain board minutes. Fine. Then let a jury sort out those contradictions. As a judge ruling on a summary judgment motion, Kimball's not allowed to make determinations about credibility.
Similarly, Judge Kimball appears to have discounted then-Novell-chief-negotiator Chatlos's credibility because, as the judge noted, Chatlos's wife currently works for SCO. Okay, I can see why you might reach that (rather cynical) conclusion -- but only if you're a juror.
Kimball also appeared to discount paralegal Madsen's testimony because she was a lowly paralegal, whereas Bradford was a general counsel and Braham a partner at a big firm. Fine, but only if you're a juror.
Were there any conceivable reasons why a juror might have chosen to give former-Novell-general-counsel Bradford -- who, Judge Kimball says, "oversaw the negotiation and drafting" of the contract -- less credence than Judge Kimball chose to? Well, come to think of it, there was. Though he doesn't mention it, Judge Kimball did have before him a May 17, 2007, declaration from Lee Johnson, a longtime friend of Bradford's, who swore that he had repeatedly asked Bradford about the 1995 contract after the dispute first arose. "Without exception," Johnson wrote, "[Bradford] has told me that he was not significantly involved in that transaction and did not know much about it. . . . I even . . . specifically recall asking, 'How could you not know who owns the copyrights given your position at Novell at the time?' David's response was . . . : 'I was busy on more important things and was not involved in that level of detail in the Santa Cruz transaction.'
"The Johnson affidavit then goes on to allege that, after Johnson found out about Bradford's testimony in the case -- the testimony that Judge Kimball ultimately so heavily relied upon -- Johnson confronted Bradford about the apparent shift in his recollection. "In response," Johnson wrote, "he left me a voice message where he specifically stated that he did not remember any of this but he had gone back and read the agreements a few times and concluded that this must have been what happened. In other words, he does not remember this at all but has now convinced himself that this is what happened."
Would a juror be entitled to disregard Johnson's testimony? Of course. Maybe Johnson's a crank. Maybe he has some sort of pecuniary interest in the case. And why didn't Johnson save the voicemail message if this really happened? Those are all good questions for the jury to ask and answer.
It's not hard to predict what any appellate lawyer hired to defend Judge Kimball's ruling would say in its defense. He'd argue that, notwithstanding some unnecessary verbiage, all Judge Kimball really did was decide that the contract itself was so unamibuous on its face that all of the "extrinsic evidence" -- the testimony from the CEOs, the other negotiators, Bradford, Braham, Johnson, Madsen -- was all irrelevant. Kimball does, in fact, reach that conclusion in the end. Since the contract language was not in question, and not reasonably susceptible to any other interpretation than the one Novell now gives it, Judge Kimball ruled, he could properly grant Novell summary judgment.
But that argument is wrong for at least two reasons: First, from the opinion it's apparent that Kimball didn't have the foggiest idea what the parties were up to with this contract -- and who could blame him? -- until he waded deeply into all the extrinsic evidence. Having done so, he then decided that Bradford's and Braham's testimony seemed to make the most sense over all. But having used the extrinsic evidence to figure out a plausible hypothesis for what the contract might mean, he can't then proclaim the contract to have been "unambiguous" all along, entitling him to disregard all the diametrically conflicting evidence he had to discount and discard in the process of reaching his conclusion. That's like using a rope ladder to climb up into a helicopter, and then pulling up the ladder so no one else can climb up after you. (Okay, not a perfect metaphor, but I think you see what I'm getting at.)
Second, the language of the contract is ambiguous. What it gives in one provision, it takes back in the next. It's not fully consistent with either party's claims, and never will be.The asset purchase agreement says that Novell sold to Santa Cruz "all rights and ownership of UNIX ... including source code . . ., such assets to include without limitation" a long list of specific products. SCO argues (and presented an industry expert who said) that if you're buying "all rights and ownership" of a software business "without limitation," you're obviously buying the copyrights. This contractual language is also inconsistent on its face with Novell's claim that it was only selling a license to use UNIX for limited purposes.
On the other hand, Novell rightly points to another page of the contact, which lists five categories of assets that are to be "excluded" from the deal. Three of the first four categories concern NetWare products -- a software business that Novell was unquestionably retaining control of -- while the fifth says "all copyrights and trademarks, except for the trademarks UNIX and UnixWare." SCO claims that "all copyrights" here was supposed to mean "all NetWare copyrights," and that the rest of the contract would make no sense if Novell was also retaining the UNIX copyrights, too. Novell, on the other hand, says "all copyrights" means "all copyrights," and, therefore, we just don't need to worry about what "all rights and ownership of UNIX . . . without limitation" could have meant on the earlier page.
Looks like ambiguity to me, confirmed in spades by the highly conflicting extrinsic evidence that Judge Kimball allowed into evidence and then laboriously waded into and picked his way through before deciding in the end that, you know what?: This contract really has only one possible interpretation!
Readers may have long ago wondered why I'm getting so worked up about this. After all, you may be thinking, if Kimball's ruling is really as bad as I say, won't it just get reversed on appeal? Well, that's the thing. SCO's got about $10 million in cash and it's burn rate seems to be about $1 million per quarter. It's not just fighting Novell and IBM, it's fighting the clock. Kimball's ruling could be the coup de grâce. (On Friday Judge Kimball squelched SCO's long-shot attempt to seek an immediate appeal of his August 10 ruling, so SCO will need to wait until the trial is complete before it can start the appeals process.)
Litigants, especially unpopular ones, are entitled to their day in court. In this case, the last word on SCO's controversial claims should have been delivered by a jury, not by Dale Kimball.