MS screws-up video remake, and admits it wasn't real anyway
Phrases like 'blatant attempt to mislead the court' seem to spring to mind...
Microsoft's fatal video demo seems to have finally vanished beneath the waves yesterday, as the company failed to remake it adequately. To make matters worse, the company also confessed that the Internet connectivity in the earlier video hadn't been real in the first place. In principle that admission needn't necessarily have been surprising - Internet connections are so variable that it's practically impossible to produce a level playing field for a 'scientific' test. Simulating a connection is therefore a far more sensible route to take, and we're pretty sure we saw a Microsoft reference to Allchin's original tests as being based on simulated connections earlier this week. If anybody can remind us where we saw this, incidentally, it would be a big help, as right now we can't seem to lay hands on it. But even if Allchin's tests were perfectly legitimate in using simulations, the video Microsoft showed at the beginning of the week did not make that clear. The reverse, in fact. It was originally presented as a video of a duplication of Allchin's tests, and as our transcript makes clear, anyone watching it would come away with the impression that it was showing a real live connection to the Internet. On numerous occasions Yusuf Mehdi's voiceover specifically states that the computer is connecting to the Internet. For example: "When I click on that icon, we'll connect to the Internet … We are now on the Internet … we're actually connecting out to the Internet and fetching that data." Enough? There's plenty more. But yesterday a Microsoft spokesman was telling reporters Microsoft was "using computers in a studio to illustrate the points that we had discovered in the laboratory," which as far as we can make out means there was no live connection, or at least no continuous live connection as shown/implied by the video. So take a look at what has happened to the video over the past few days. On Monday it was a duplication of tests already conducted by Allchin, using "virgin" machines. It was scientific, and accurate. The discrepancies in the video having been shown by the DoJ, it then moved to being a still scientific and accurate demonstration where unfortunately there had been a couple of minor and irrelevant changes made to the Registry on one of the machines. Then by yesterday it wasn't scientific and accurate at all - it was an 'illustration' of Allchin's points, a TV commercial, you could say. You could also say, based on the transcript, that it looks awfully like a deliberate attempt to mislead the court. These problems are compounded by the failure to duplicate the tests on Wednesday night. Microsoft had won permission to try for a second, clean video on Wednesday, but that evening - surprise, surprise - it found it impossible to achieve consistent bandwidths on the Internet connections. As we all know this already (it was where we came in, wasn't it?), it's difficult to understand why, on Wednesday afternoon, Microsoft was under the impression it could go to an office, set up some computers and achieve consistent connections to the Internet, so it could perform scientific tests. Are they just stupid, or is there another explanation? ® Related Stories: Full video transcript Allchin heads for video remix MS exec recants over video inconsistency DoJ skewers MS exec over falsified video Complete Register trial coverage
Sponsored: Beyond the Data Frontier