The MS campaign to rein-in Intel multimedia development
Engstrom admits MS need to "own all the driver software down to the metal"
MS on Trial In the Microsoft trial, Eric Engstrom is being used to challenge the account of Microsoft's Intel relationship, as told by Steve McGeady and Intel's documents. Strangely, McGeady's name was never mentioned: he has evidently become a non-person. An Intel internal email from Gerald Holzhammer on 13 April 1995 summarised a face-to-face meeting with Engstrom, Carl Stork and Marshall Brumer of Microsoft. He concluded that Microsoft wanted to own the drivers in Windows 98; that nobody but Microsoft was qualified to do good driver software; that Microsoft would not collaborate on NSP; and that Microsoft had "completely missed the boat" on developing a compelling state of the art media subsystem for Windows 95", according to a confession by Carl Stork. Engstrom sent a status report email on 26 May 1997 about the "collaborative" multimedia efforts of Microsoft and Intel. It was very revealing: "The goal for our interactions with the Intel Architecture Labs is that all their efforts are neutral or positive for our strategic mm [multimedia] initiatives. I think it is unlikely that we will achieve 100 per cent of this goal". It turned out that Microsoft was obsessed at this particular time with collaboration between Sun and Intel over Java-based competing multimedia standards - the MPEG4 specification. Engstrom's goal: "Intel to stop helping Sun create Java Multimedia APIs, especially ones that run well (ie native applications) on Windows". Engstrom was unhappy that every IAL partner slide referred to Sun. His proposed actions included Microsoft stopping giving any help to Javasoft wherever Microsoft agrees to ship Intel technology in DirectX media. Microsoft was concerned at Intel's efforts "to do multimedia on Intel-based Windows platforms" such as RSX (3D sound), RD (sprite engine), video playback extensions, and procedural textures generally. So far as large model management was concerned, this was competitive with "our" HP Jupiter technology. Engstrom ruefully noted that Intel had "a rather large evangelist force that is willing to use monetary incentives." Whatever next? It could not have helped Microsoft that Chrome (the multimedia browser) was projected to Intel as "the big carrot", in view of what subsequently happened to Chrome. Engstrom said in his testimony that Intel wanted some software that would produce special effects to make its processors look good, but Microsoft just pressed on with pressuring Intel to stop software development. Rather than admit that Sun and others were important Intel partners, Engstrom puts Intel's behaviour down to Microsoft "not paying enough attention to Intel" rather than recognising that Microsoft's lack of expertise and arrogance might be part of the problem. Microsoft needed "to own all the driver software down to the metal" Engstrom agreed, because in 1995 "we were trying to move the games market forward. They had been basically stuck using very similar technologies for ten years because all of the advances in Windows hadn't helped games at all. So, the purpose of the DirectX technology was to make the Windows environment and Windows hardware acceleration accessible to games. At that time, given the speed of the processors and the hardware state, you wanted the game to know exactly what was happening. If there was a bunch of code running in there that the application didn't know about and couldn't test against, you would get bizarre behaviour in the game. It would stop being realtime." Surely there is no relationship between Intel's friendliness towards Gates' arch-enemy Sun and Microsoft's violent reaction against Intel producing software that Sun could use? ® Complete Register trial coverage
Sponsored: Beyond the Data Frontier