Alex headshot

AlBlue’s Blog

Macs, Modularity and More

A week since iDay

2005, mac

It's been almost a week since Apple dropped its bombshell that it it is migrating to Intel processors at WWDC 2005. What is perhaps more surprising than the fact that it happened is how that it happened.

Although NEXTSTEP worked fine on multiple operating systems (the Motarola 68040 as well as x86), the transition to Intel platforms effectively killed NeXT. Prior to the release on Intel hardware, NeXT made its own expensive, black, hardware (the pizza-slap NeXTstation and the NeXTCube) that ran its proprietary operating system on. However, in a bid to get more customers onto its platform, Steve Jobs announced a version available on Intel, and shortly afterwards, the NeXT hardware collapsed and they turned into a software-only company. Fortunately for NeXT, Steve Jobs, Apple and numerous Mac fans, Apple bought out NeXT and Mac OS X is the fruit of that forbidden love.

So is history repeating itself? It certainly seems that way from the outside. At WWDC, Steve announced that all computers post 2006/7 would have Intel Inside (da da da da ding) -- effectively killing off interest in the current range of PPC hardware. It's even known as the Osborne Effect. Admitedly, he said that there was some cool stuff coming up (well, certainly not a G5 then -- there's no way that is a cool processor, even with that many fans) -- but unless you're completely engulfed in the Steve RDF, you aren't going to fall for that one ...

What's more interesting is why this has happened. Clearly, IBM's got a lot of grunt (although Apple is still a small customer) and especially since the Cell processor was announed for the PS3 and XBox 3. Any dreams of a Mac wth a Cell on board are unlikely to happen in the near future; and in any case, the PPC that's used in the G5 had a bunch of extras (e.g. AltiVec) that aren't present in IBM's like of Power5 hardware. This could well be a case of Steve Jobs reacting emotionally to a perceived bad deal (Disney, iCon ...) and throwing a tantrum. It might turn out to be the most expensive tantrum of all.

But the future holds some certain amount of interest for the Mac faithful. What's going to be key is to limit the size of the number of copies that Mac OS X breaks into the wild on out-of-the-box hardware. If OS X ever gets to run on a beige box, then Apple might as well kiss its hardware sales -- at least, from a desktop point of view -- goodbye. On the plus side though, an unsanctioned leak might help to create interest in the OS and win more Microsoft switchers than it would otherwise do.

Personally, I think the future is multi-core, not faster chips. CPUs can only go so fast for a particular power/heat outage, and although shrinking can help solve that, it will run up to a limit. However, a PC with dual chips (or even quad chips) would be twice as fast, and probably a similar amount of power.

Can it be the case that Intel has a better roadmap than IBM? Or is that just a cover story? Well, one thing that might make it true is if Intel were launching a range of low-power dual-core chips; ideal for a PowerBook. Laptops aren't going to go commodity, and as anyone who has used a PowerBook knows, Apple design these things well first and foremost. A laptop with effectively dual processors? It's an interesting idea; though of course, the iBooks would have a single processor.

But Apple may have terminally shot itself in the foot before these plans even get off the ground. With the migration likely to affect many developers in terms of support and maintenance, they may just give up on the older PPC system. Others, especially those who still run OS 9 apps, are not going to change -- and the short-term demand might even stay the same as these users hoard G5-class machines. The biggest loss, however, is in the scientific community who use G5s not for their processor speed, but for the AltiVec instructions -- the parallel processing register that's optimised for the likes of Final Cut Pro and Photoshop. (That's why PhotoShop is repeatedly shown in tests between two machines; it can take advantage of this parallel processing unit.) The other major loss is the 64-bit architecture; there doesn't seem to be any support announced at present, so we're going backwards into the realms of a 32-bit architecture. Whilst this almost certainly makes sense for low-power devices such as the PowerBooks, it doesn't make sense for the PowerMac or Xserve systems that potentially need to use a lot of memory.

Who are left as customers? Well, there's always the geeky 'have' crowd. But I will probably stick with10.4 on my PPC machines at home; when 10.5 comes out, mabye I'll upgrade them then; but if 10.6 is intel-only, that's the end of the line for my machines at home. Would I upgrade? Not until 2010; the Macs that I've bought in the past have tended to last at least 5 years.

Of course, if this gamble pays off, then Apple might just end up increasing its market share in the current crop of users, but it's a hell of a gamble to pay. With the next several months of sales potentially evaporated, Apple's going to have to come out with some really cool toys like a media centre or a video iPod to stay in the market. The only way to find out is to wait it out; but for an advance preview, you can see which way it's sliding on the Apple Stock price ticker.