How MS tried to dig itself out of video hell
Unfortunately, demo two seemed to be full of legal holes as well...
MS on Trial The DoJ had arranged that Felten, his two former students, and some DoJ lawyers were at the Washington offices of Sullivan and Cromwell at the appointed hour of 8.30pm last Wednesday for the videotaping of the demonstration. They were kept in the ground floor lobby for some two hours, and denied access to the sixth floor conference room where the demonstration was to be held. It is coincidental that the conference room overlooks the White House, and there is no independent confirmation that the occupants were watching for Monica, who was in town on another matter. Allchin claimed he had not been told that the DoJ team was waiting, and that there had been a problem with the three camera crews. He had been deliberately kept in the dark, it seemed, so that he wouldn't be able to answer searching questions about the delay and the logistics in the morning. It is a well-known ploy to keep opponents hanging around late at night, in the hope that their concentration is diminished. It transpired that Microsoft went out and bought six Thinkpads from a retailer, and some software. Allchin said he did not think whether these machines would have Prodigy pre-loaded. Allchin, who has a PhD in computer science and many years experience as a senior executive at Microsoft in charge of operating systems development, evidently did not consider using the find files or folders facility to check this. Allchin said the next day that he "started learning the IBM machine, and that it took him a great deal of time to establish a connection. Allchin told Holley: "We actually put quite a few ISPs in there because the phone situation was a true mess" (this in the conference room of very expensive lawyers). The boxes containing the Thinkpads were opened and the machines set up without the DoJ party being present. In view of the false claims made about the first Felten videos, not allowing the opening and set-up of the Thinkpads to be witnessed must invalidate the integrity of the demonstration completely. Allchin was accompanied by at least one of the staff who had been involved in the original demonstrations, so there can be no certainty that the present demonstration was an honest one, as the machines could have been preconditioned. Perhaps Allchin appeared to be embarrassed at his personal incompetence in using a PC, making connections, and his unfamiliarity with the detail of the Felten experiment. Allchin claimed to Boies: "The web sites dealing with the Deluxe CD go up and down, so if you take a regular Windows 98 system, occasionally you will get that same error. But I had a developer go through the code and determine that there were unique characteristics that happened in this particular case. So, it is possible that if you're at home with this Deluxe CD player that you could get an error like this if the web site was down." This detracts further from any remaining integrity in the demonstration. Allchin did not try to demonstrate the previously claimed performance degradation during the Windows update. When pressed by Boies about this, Allchin said: "I had talked to individuals, developers, who had seen performance degradation on their machines. This is out of the lab. In other words, just in use. I think I testified that I personally hadn't seen that. "I had not personally seen that when I used it, and people told me that I hadn't been using Dr Felten's program and doing enough things long enough, but I personally never saw that. And what I said on the tape is what I believe, which is that the only way, being the engineer that I am, that I would trust this is if I set up two machines, a client and a server in, you know, unconnected to the Internet in any way, and run a very confined test, which we have done, and in that confined test we saw the APIs to urlmon.dll and to mshtml.dll. Those interfaces we saw performance degradations on the test that we did. Although I didn't see it in day-to-day use, I believed that you could see those degradations if you were running things like what are called web crawlers and other things that are accessing URL's a lot. What those do is walk around the web and gather pages if you're accessing that code a lot. But I didn't attempt to do it because I didn't believe it was scientific ...". This confession, that Allchin had not been present to see the alleged degradation test, makes his evidence hearsay, and highly suspect. Boies came back to the whole point of the Felten work and asked Allchin whether he agreed that the program was not a commercial product, had been designed to demonstrate a concept, and might have problems. "Yes," he replied. Allchin confirmed that in the normal course of software engineering, when you identify a bug, you try to develop a bug fix, and then you test it. Microsoft had not tried this with Felten's program. Boies continued: "Just focussing on Microsoft's products for a moment, is it common or ordinary in early versions of programs that Microsoft distributes, including parts of Windows 98, for there to be problems with performance, slowdowns or bugs or applications breaking? Allchin did tell the truth: "Well, we -- yes, software does have bugs in it. We would all like there not to be any, but there are." With this, Allchin was sent back to Washington state and Judge Jackson announced that he would like to see counsel in his chambers. The reason for the judge's unusual move has not been disclosed, but it is a reasonable speculation that Judge Jackson wanted to air his concern at the credibility of Microsoft's witnesses, and warn Microsoft's lawyers about their future conduct. It is inevitable that this episode will cause him to reflect on the credibility of other contentious Microsoft evidence. ® Complete Register trial coverage
Sponsored: Benefits from the lessons learned in HPC