Apple to move on Merced
Click here for the previous page IBM, meanwhile, continues to see the chip in multi-processor servers, the role that its Power chip-set, which formed the basis for the PowerPC, was developed for. That's why it wasn't too keen on AltiVec, a technology far less relevant to the kind of systems IBM is interested in. IBM reckoned the silicon spent on the AltiVec engine could be better used for other, more server-oriented functionality. The disagreement between IBM and Motorola over AltiVec was the first sign that the AIM alliance might someday come undone. Intel has a stake in seeing that happen, and might well chose to encourage Apple to believe it's better off with Intel. There's nothing underhand here -- Intel is simply pursuing another customer. And don't forget Motorola is suing Intel for allegedly poaching a number of PowerPC engineers, so Intel has even more reason for making PowerPC development appear in crisis. However, rumours of AIM's death are probably exaggerated. Late last year, Mike Attardo, general manager of IBM's Microelectronics Division, said the two companies will soon come to a solution to their disagreement. As he put it: "The three companies are motivated to co-operate, and we will co-operate." The relationship has been going through some turbulence of late -- what relationships don't? -- but there's no real sign of a major falling out between any or all of the partners. And even if IBM did call it quits, Motorola and Apple have more than enough motivation to stick with their version of the platform. So why should Apple be working with Intel at all? Why, if The Register's source is over-egging the pudding on the PowerPC side not be doing the same with MacOS X? A quick look at what both companies are currently up to shows that there's some important stuff here. Apple first. The idea of a largely-platform independent OS has been doing the rounds at Apple for ages. As soon as it began work on the Pink OS (the one that later became Taligent) and began messing around with microkernels, platform independence has been an issue. Perhaps not always part of the core strategy, but present nonetheless. Platform-independence appeals to Apple because it allows you to get your OS out to the widest possible audience, even those who don't want to ditch their existing, non-Mac hardware. Apple extolled a multi-platform future for the MacOS shortly after it bought NeXT, claiming that the Yellow Box API that formed the heart of what was then called Rhapsody would run natively on Intel and, separately, under Windows. Intel compatibility has largely been ignored since then, but with MacOS X Server based on a variant of the Mach microkernel and cross-platform BSD Unix, it can easily be moved from processor to processor should Apple feel the need to do so. And it probably does feel that need. Apple wants MacOS X Server to be something more than an OS for Mac workgroups, and that means competing with Windows 2000 Server (aka NT). Windows 2000 will offer far more to corporate network administrators than MacOS X Server does, but that makes it increasingly unappealing in the small-to-medium enterprise arena. It's very late, has huge question marks floating over its ability to deal with the Millennium Bug and adds so many new features, it's going to take a lot of time for IT departments to fully evaluate it, let alone implement the thing. Click here for the next page
Sponsored: RAID: End of an era?