Schmalensee trips on DoJ banana skins

Government lawyer lays his traps, and MS' witness walks right in

MIT economist Richard Schmalensee continued his Microsoft consultancy, answering questions from David Boies for the DoJ. His desire to please Microsoft showed all the fervour of a convert to a cause, so that he appeared to be more of a Microsoft salesman than an expert witness, abandoning any scientific rigour from his discipline in defending his client. Schmalensee suggested there was a time when Netscape and Microsoft were friends: there was "a period when Microsoft was treating Netscape as a preferred, indeed one of its most preferred ISVs. It was extending special co-operation and so forth". His best example of this alleged friendliness was "a presentation listing of hot Internet applications made by a Microsoft person in Germany in which Netscape was featured [in mid-1995]". This is indeed stretching the bonds of friendship... There was the expected verbal jousting about the term 'browser' versus 'browsing functionality', with Schmalensee still maintaining the Microsoft line, which has been essentially blown away in previous evidence from Microsoft emails, that developing a browser and making browsing part of Windows was a simultaneous and early decision: "I know someone was assigned the task of writing a browser. They had difficulty doing it. They licensed code. They may have been uncertain, when they decided to start that group, what the ultimate design decision was. That was 93 or early 94. The technology was in flux. Windows 95 was a bit of a gleam at this stage." Schmalensee then said this was in the first half of 1994, and that there were "roughly contemporaneous discussions, again with Booklink and then with Spyglass, having to do with possibly licensing code". Of course it is understandable that Schmalensee would want to help Microsoft in its rewriting of history, but the facts are different. What actually happened was that Microsoft had to license Internet Assistant from Booklink Technologies, so that Word could convert documents to HTML. Microsoft tried to buy Booklink's browser for $2 million, with no royalties, but was turned down. Unfortunately for Microsoft, just after Microsoft did the licensing deal, Booklink Technologies was bought by AOL, and had Bill Gates known this was going to happen, he certainly wouldn't have hyped the product so much at Comdex in November 1994. It was no doubt with some pleasure that AOL trumpeted that "America Online Booklink Unit announces licensing agreement with Microsoft". That rather firmly places the date at the end of 1994, and not in the first half at all. The Microsoft-sanitised version of events by Schmalensee omits that in the Fall of 1994, Microsoft had tried to license Navigator. The first public revelation of this was on 27 September 1995 in a Dow Jones story. It was for this reason that Microsoft turned to Spyglass. Although Schmalensee seemed unaware of it, the Spyglass negotiation was signed on 13 January 1995, after Microsoft had previously turned down Spyglass on the ground that it was going to build its own browser. Microsoft had tried to go around Spyglass, holder of the master licence from the National Center for Supercomputing at the University of Illinois, and deal directly with NCSC. It wasn't necessary, because Microsoft out-negotiated Spyglass completely, and got away with a deal that soon left Spyglass struggling. The deal was for the Mosaic browser, which Microsoft would call Internet Explorer. "The original code [for the browser] has been tightly integrated into the operating system," said Naveen Jain [Microsoft's senior technical manager for MSN] -- the implication at the time being that it would be difficult to remove the code if the DoJ objected to Microsoft bundling Windows 95 with MSN, which was a serious possibility shortly before Windows 95 was released. It seems that Microsoft had missed Joe Garvey's piece in Inter@active on 20 June 1995 about this, or it would have been shouting it from the rooftops. How "tightly integrated" Windows 95 and IE were at the time is a matter of conjecture, but Microsoft has produced no evidence that there was code sharing. It is noteworthy that accounts of these events usually omit that most of the coding for Mosaic was done by Eric Bina rather than Marc Andreessen (later of Netscape), and that Bina used as his basis the code by Tim (Rule Britannia) Berners-Lee. Another fact that seems to have escaped the contestants is that Microsoft extended its contract with Spyglass on 7 December 1995 (and added a Web server compatibility agreement), a few hours before Gates' announcement that Microsoft was giving its official recognition to the Internet. Spyglass' licensing arrangement with Microsoft ended in December 1998. So far there appears to be no hard evidence as to when Microsoft actually decided to give IE away free. Schmalensee was asked about a DoJ interrogatory (a question to which Microsoft has to provide an answer under oath): "Identify each current or former Microsoft employee who was responsible for the decision to price and distribute Internet Explorer free or without separate charge and, for each person identified, describe the role of that person in such decision and the date(s) of such decision." Microsoft's response was: "Bill Gates, Paul Maritz and others made the decision not to charge for Internet Explorer technologies apart from the price of the operating system. Bill Gates announced that decision during his Internet strategy speech on 7 December 1995." Schmalensee then put his foot in it by saying "I guess that strikes me as a non-responsive answer, since it doesn't give the date of the decision". At this point, Richard Urowsky, one of Microsoft's lawyers, bounced up to do his minder's job, and objected. His particular purpose seemed to be to convey a message to Schmalensee to be careful as to what he said. Schmalensee subsequently referred to the use of "announcement" as "a fairly difficult and slippery term", which scarcely helped his client's case. Since the Microsoft response was given under oath, the DoJ is entitled to assume that Microsoft decided to offer IE free at the late-night pizza-eating session on 6/7 December 1995 in Redmond in which the fax was kept busy agreeing the details of the Java licensing agreement with Sun. It is known that Spyglass was utterly amazed when it was announced that IE would be "forever free" because of the implications in its contract with Microsoft, which is understood, unusually, to allow for a royalty. Of course, the royalty on a price of zero is zero. Gates is believed to have a strong aversion to having royalty agreements with anybody on the grounds it would make it possible for Microsoft's claims about sales to be checked (by an auditor if necessary), which could prove embarrassing and costly for Microsoft. This is why Microsoft settled out-of-court when Spyglass eventually demanded an audit of IE sales following Microsoft's claim for millions of copies of its browser being deployed but no money from Microsoft. At the same time, this provides some circumstantial evidence that Microsoft had decided that IE would be free some time earlier, and did not therefore mind allowing Spyglass to have a royalty that it knew would be zero. Boies couldn't resist taunting Schmalensee about his colleague Professor Michael Dertouzos, Director of the MIT Laboratory for Computer Science, who was going to testify. Bill Neukom, Microsoft's head lawyer, had said on 4 September 1998 in a press release that he would testify on the "technological benefits of Microsoft's decision to build Internet features into Windows". In the gentlest possible way, Boies asked Schmalensee if he had read Dertouzos' deposition and "if he described the browser as an application" in his deposition. In fact Microsoft dropped Dertouzos because he had described IE as an application, which impoverished Dertouzos mightily as a result, since he did not receive much of the Microsoft shilling. It was, of course, just coincidental that the MIT Technology Review, in its current edition, has an article Programs to the people by Charles Mann dealing with free software, which mentions how "Goliath gets nervous" (about Linux) and ends with a quote from Gandhi: "First they ignore you. Then they laugh at you. Then they fight you. Then you win." Schmalensee was tempted into venturing further opinions about software that he was unqualified to make. The theme started with how an application might or might not alter the operating system, and progressed to whether Word would cease to be a separate product if it were combined in the operating system. The breakfast cereal expert waffled on logorrhoeically (think of it as verbal diarrhoea) and finally came out with a startling scenario: "Suppose I am a WordPerfect user. Microsoft does what you describe [combines Word with Windows]. And I prefer WordPerfect to Word. And I find that the convenience of using WordPerfect, the speed, my ability to install it -- that in some obvious way I can't do as well as I could have done before, that's a harm." Techno-sabotage is a well-documented anti-competitive technique by Microsoft, which makes one wonder if Schmalensee had been privy to some Microsoft plan to take action against WordPerfect. On the face of it, it seems unlikely that Microsoft would merge Word into Windows, because of the very significant revenue stream it receives from the Office suite. Boies continued laying a trail of banana skins, asking Schmalensee an innocent question and then watching him slip on each as Boies produced quotations from writings and earlier testimony that controverted the present testimony. More and more it seemed as though the court was hearing evidence-to-order, and not true expert opinion. Urowsky became increasingly anxious, frequently jumping in vain up to object, as Schmalensee was systematically demolished as an expert with intellectual integrity. One example concerned tie-ins. Schmalensee had testified in the Data General case that there was a tie: "Power over price... includes the power to impose a tie, or other burdensome terms, in lieu of charging a high dollar price," he had said. There was also a discussion of software lock-in, and switching costs when changing systems. Boies proffered a banana skin, asking if one of the purposes of combining IE and the operating system was to increase the distribution of IE. "It certainly had that effect... I've no reason to think it was a surprise or undesirable to Microsoft." Then another Boies banana: "Do you recognize that one of the effects of combining IE with the operating system was to make it more difficult for Netscape to distribute its browser?" "I think that the answer is 'yes', " said the slippery one. Proffering another banana, Boies asked if combining IE with Windows made it more difficult for Netscape to become a platform competitor for Microsoft. Schmalensee: "I think that the fact that Microsoft offered a competitive browser and distributed it widely did make it more difficult..." Boies then turned to the question of the intent of a company in anticompetitive actions, and when Schmalensee was not forthcoming, Boies dug out the Bell Atlantic case in which he had testified that a pattern of conduct could be predatory, and therefore anticompetitive. Cunningly, Boies made use of the case a second time, by leading into how Microsoft had failed to fulfill its obligations with respect to Java. This time, Schmalensee stood on the banana skin, but Judge Jackson may well have thought that he slipped because of the analogy between AT&T's predatory actions that violated antitrust laws, and Microsoft's actions towards Java. Having established a virtual skating rink with all the banana skins, Boies then tossed to Schmalensee some technical questions about shared code in IE 1, 2, 3 and 4, the DLLs involved, and code sharing. It was a disaster for the skater: he shouldn't have been in the banana rink trying to answer computer science questions at all, and he landed up in a very unflattering position. Boies produced a Microsoft email from the Bristol versus Microsoft case (in which Schmalensee had testified): "We are now describing IE as part of the OS and putting it in the box. Conclusion: they already have rights to things like IE. It's a legal stretch, but I want to hear from our attorneys and you, Bob. But in any case, I'd rather use IE as a bargaining chip to clarify the extent of their rights; ie., treat IE as a separate deal and, thereby, set the precedent that apps in the box aren't really part of the OS." It now seemed that even Microsoft was willing to regard IE as something separate from Windows if it suited them. It was a very significant document. Urowsky was objecting furiously, but was overruled by Judge Jackson, who paraphrased Boies argument very well, thereby showing he was paying very close attention. The date of the email was 22 December 1995, just a couple of weeks after Microsoft's Internet strategy meeting in New York. "We have here Microsoft talking about applications in the box not really being part of the OS," Boies pointed out. Schmalensee valiantly tried to pass this off as "one person in Microsoft" putting forward a view, but he slipped again. Boies said he was concerned that Schmalensee says: "Pre-installation of software by Microsoft on the Windows desktop is not a significant method of software distribution by Microsoft or anyone else." Admittedly, Schmalensee is described as an economics expert, but this opinion certainly shows he has no knowledge or sound views about the software industry. Although Boies had by this time unexpectedly used up his supply of bananas, it was not hard to find evidence to the contrary. A Microsoft OEM marketing review of 27 May 1998 said: "'It came with my computer' is the number one reason people switched to IE." Even worse for Microsoft was: "Conclusion: OEMs are the best vehicle to gain browser share," which is pretty close to admitting that IE is a separate product. When Schmalensee did not know something on one occasion, he admitted that NERA staff had done the work, and that he had not read a deposition quoted in his direct testimony. Professionally, the weakest part of Schmalensee's direct testimony was the part dealing with survey data commissioned by Microsoft to investigate browser share. It turned out that the sample size was less than 300, and that the results were in sharp contrast with those from surveys using much bigger data sets. It was an attempt to show that IE market share was lower than it probably was. By this time the banana skins were beginning to get pretty smelly. ® Complete Register trial coverage

Sponsored: Designing and building an open ITOA architecture