audio | update WASHINGTON--Microsoft disavowed a videotape segment it played during its antitrust trial under way here, acknowledging that the exhibit should not be viewed as a literal depiction of a computer test the company introduced to undercut the government's case.

Microsoft spokesman Mark Murray said the videotape was merely an illustration, not a demonstration, of laboratory tests conducted to show that Windows 98 ran more slowly when its Internet-browsing capability was turned off.

The tape used "computers in a studio to illustrate the points that we had discovered in the laboratory," Murray told reporters.

The company also acknowledged that a new test it performed overnight to try to rehabilitate the "bollixed" videotape did not fully resolve the question of whether disabling Windows 98's Internet technology degrades the operating system's performance.

That is a key issue in the government's case, which accuses Microsoft of welding Internet Explorer onto Windows 98 to foil competition from Netscape Communications. Microsoft says it made a single, integrated product to give computer users better technology, not to stifle competition.

Monday, Microsoft Senior Vice President James Allchin referred to the original videotape demonstration to debunk testimony from government technical expert witness Edward Felten, who had said he wrote a program that disabled Internet Explorer without hurting Windows.

A new 70-minute tape, produced Wednesday night after the government pointed out problems with the first version of Microsoft's video, featured Allchin conducting tests on a version of Windows that was running with Felten's program.

More coverage on CNET Radio

It showed that some software that was dependent on browsing functions didn't work when the Felten program was running. The new video, though, did not support Allchin's contention in written testimony that Windows' performance was slowed 200 percent to 300 percent by Felten's program.

"It would be incredibly difficult for me to show the performance issues. That has to be in a controlled lab environment. We don't have that here," Allchin testified.

Yesterday, Microsoft shares fell 7.75 to close at 159.0625. In mid-afternoon trade today the stock was down 1.6875 at 157.375. Analysts said the stock had been overvalued by a recent run-up and said they doubted yesterday's drop was due to the trial developments.

"Most people are taking [the trial] with a grain of salt," said John Puricelli of AG Edwards.

Ill-fated tape
Because Microsoft's latest video didn't attempt to test why Windows ran more slowly with Felten's program, "the tape demonstration didn't have a lot of relevance," said David Boies, the government's chief trial lawyer.

After the trial adjourned for the week, Murray acknowledged that the ill-fated tape segment was not an actual demonstration of lab tests, as Allchin had said in court. The narrator in the original videotape "knew in laboratory conditions performance was degraded" and "he commented on that," Murray told reporters.

During Allchin's testimony Monday, the videotape was described by Microsoft lawyer Steven Holley as "a videotape of a series of demonstrations done under Mr. Allchin's supervision."

The new tests were conducted to allay doubts voiced yesterday by U.S. District Judge Thomas Penfield Jackson about the accuracy of the first video. He granted Microsoft's request to redo the tests after Boies exposed the video's discrepancies. Holley told the judge the exhibit "got bollixed up."

Jackson said yesterday in a private conference with lawyers that, while he did not believe Microsoft set out to deceive him, Boies had done "a very professional job of discrediting those tapes."

Different speeds
Allchin said the company couldn't reproduce the laboratory results shown in the first videotape because Microsoft technicians had to depend on Washington local phone lines to make Internet connections needed for the tests.

"I didn't attempt to do that because I didn't believe it was scientific," Allchin told Boies.

During the Wednesday night tests, several different online connections were made to the Internet, at widely varying speeds.

Microsoft executive Tod Nielsen argued that at any rate, the Internet speed issue was not the primary thrust of the video demonstration. He said the new test showed that the government witness' browser-removal program "doesn't remove Internet Explorer and breaks Windows.... The [Internet] performance was an additional thing that it did."

Earlier in the day, Murray had told reporters that the new test "resolves once and for all any questions about the reliability of our evidence."

Copyright 1999, Bloomberg L.P. All Rights Reserved.

Related News
• Judge: Video discrepancy "very troubling" February 3, 1999
• Microsoft tries to explain trial glitch February 3, 1999
• Get this story's "Big Picture"