By Siobhan Kennedy The Judge presiding over the Microsoft antitrust case in Washington said yesterday he found inaccuracies in the company’s videotaped evidence very troubling and said they cast doubts on the entire tape’s credibility as evidence. Speaking after the government’s lead attorney had just revealed no less than five additional errors in the software […]
By Siobhan Kennedy
The Judge presiding over the Microsoft antitrust case in Washington said yesterday he found inaccuracies in the company’s videotaped evidence very troubling and said they cast doubts on the entire tape’s credibility as evidence. Speaking after the government’s lead attorney had just revealed no less than five additional errors in the software giant’s video presentation, Judge Thomas Penfield Jackson said: This certainly casts doubts on the reliability of the entire videotaped deposition. Directing the court to a series of still screen shots, lead government attorney, David Boies showed that the number of icons on the screen, as well as the title bars changed from shot to shot. This, he said implied that Microsoft’s technicians had used different computers to carry out the tests, and not a single virgin machine, running just Windows 98, as the company had originally stated. The revelations forced Microsoft’s senior VP and lead technical expert, Jim Allchin, to admit that yes, there were multiple machines involved. It’s not in our labs, he tried to explain, we’re not trying to be incredibly precise. Aghast, the Judge asked him: How can I rely on it if you can’t tell me it’s the same machine? It’s very troubling. Allchin replied: I still stand by my testimony that this program exhibits these failures. I’d feel much happier if you had made the tests yourself, Judge Jackson replied. I did make the tests, but it’s not what I’m seeing here, a humiliated Allchin responded. The revelations brought the afternoon’s session to an abrupt end and after private consultation between the Judge and both parties, Microsoft spokespeople announced that Allchin would repeat the experiments, Wednesday night, in front of DOJ witnesses and present the results Thursday morning when the court reconvenes. One of the great things about the software business is that if there are some bugs in a first version of a product you can go back and fix them, Microsoft spokesperson Mark Murray joked during a frenzied press conference minutes later, Video 1.0 apparently had a few things that became confusing so both sides agree we will be doing Video 2.0 overnight. Although Allchin had offered to carry out the tests, live, in front of the judge, Murray said court rules would not permit such a demonstration. In true Microsoft form, the company’s general counsel attempted to make light of the startling revelations, saying that the company made very good software but didn’t make a very good video tape. But he was eventually forced to concede: It was our fault and we intend to go about it the right way this evening.
Redmond credibility takes another plunge
The news of more mistakes in the videotaped deposition dealt a serious blow to Microsoft, who, during the morning’s re-direct examination had tried to save its skin by offering what it considered to be a sound explanation for yesterday’s inaccuracies (CI No 3,590). During re-cross examination Tuesday, government attorney David Boies was able to show that still screen shots of the videotape, which purported to run the same Internet Explorer (IE) removal program as Allchin had used in his own tests, included a title bar which indicated that the program had not been used at all. In response to Microsoft’s attorney Steven Holley’s questions about the anomalies, Allchin revealed that the software giant had hurriedly flown in three of its technical staff from Seattle on Tuesday to ascertain why the test program didn’t tally with his own results. Allchin said that, during their flight to Washington, the technicians discovered that the installation and subsequent removal of software for the Prodigy on-line service caused the error, making it look as if the removal program had not been run. He said Felten’s test displays the title depending on what is a key in the registry [of the operating system], When Prodigy is uninstalled, it changes that key. It basically deletes it, so that the erroneous title bar appeared on the screen making it look as
if the Felten program wasn’t there. The original test was performed in an attempt to dismiss testimony by the DOJ’s witness, Princeton University computer scientist Edward Felten, who claimed a program that he and two students wrote, easily suppressed the Internet Explorer (IE) browser without impairing Windows’ performance. Allchin insists that, even in the light of the inaccurate video demonstration, Felten’s program cuts the speed of Windows by 200% to 300%. The issue of whether or not removing IE from Windows 98 degrades the operating system, is crucial to Microsoft’s defense. The government maintains that the browser can be easily removed, giving users the option to load alternative software if they choose. It argues that Microsoft deliberately welded its browser to the OS in order to gain more market share and drown out its main competitor, Netscape Communications Corp in the process. Microsoft, in its defense has always maintained that Windows 98 is a single, integrated product, developed to give computer users better technology and easier access to the internet, and not to crush Netscape or other competitors.