Stardate 20020107.2327

(On Screen): The reason that Apple is, yet again, trying to sell an innovative form factor for its computers is because it has nothing else to offer with which to stand out. It's long-awaited upgrade to a modern OS is being greeted by the customers as a decided mixed bag: they like the new-found stability of a real kernel, but aren't generally as happy about what was done to the UI, which is what most Mac fans list as the primary reason they prefer the Mac. Such high profile Mac users as Andrew Orlowski, Bruce Tognazzini, Henry Norr, and Dan Gillmor have all come out against the Aqua interface, sometimes with detailed reasons why they really dislike it. Several have publicly stated that for all its warts and unreliability, they are going back to MacOS 9 simply for the UI. So Apple right now is not really winning by trying to sell OSX.

And its real problem is that it can't really sell exceptional hardware, either. Its processors are badly behind the curve and have been for several years for the simple reason that Motorola hasn't been competitive. There is the famous 18-month clock stall, and the distinct slow rate since then at which Motorola has upped its clock rate. In the run-up to this show, there was mass speculation about a clock-bump, with Apple finally breaking the magic 1 GHz barrier, possibly by quite a lot. Some rumors were for processors as fast as 1.4 GHz. As we now know, it didn't happen.

It's certainly not because Apple didn't want to do it. Apple's top strategists know full well that the rhetoric about "MHz Myth" and the use of cooked benchmarks is wearing very thin. A year ago Jobs was still willing to do a side-by-side comparison of the fastest not-yet-shipping Mac (a 733 MHz) against a P4 running 1.4 GHz. By carefully choosing a few selected Photoshop filters which were optimized for the G4 PPC, and by using a version of Photoshop which had a bug in the Intel version (6.0), which ran filters perhaps 3 times slower than 5.5 did, or the 6.1 version that Adobe released a couple of months later, it was still possible to pretend that the Mac was "faster than the Pentium". But even among Mac fans that no longer flies; since then Intel has boosted its clock rate another 50% (and AMD has gone up even more) while Apple's fastest processors now are only about 25% faster than that 733. Adobe has released a version of Photoshop optimized for SSE2, which by all accounts really screams. And today, Intel announced a 2.2 GHz P4, which is based on a process shrink to 130 nM and which has more cache, and other improvements. Likewise, AMD announced yet another speed bump on the Palomino, with Thoroughbred (AMD's shrink to 130 nM) coming in maybe another three months which will also crack 2 GHz. At this point, not even a brazen liar like Jobs is willing to stand in front of a crowd and try to claim "We're faster!" It's notable that there was no side-by-side speed comparison today.

Moto is in big trouble financially; it's laid off or announced layoffs of one third of its staff and it is still bleeding money, nowhere more than in its semiconductor division. If it was performing badly (from Apple's point of view) when it was fully staffed and fully funded, then how can it perform after all the cuts it has been subjected to? Layoffs of one third to one half have got to affect how the division will perform in future, and not in a positive way.

So while rumors have been flying recently at the Mac fan sites about a G5 processor, or about G4's running at speeds comparable to the Athlon, the big disappointment today was not merely that there was no boost to 1.4 GHz (or even 1.2 GHz) but rather that there was no speed bump at all. The new iMacs use 700 and 800 MHz G4's, while there was no announced change in the PowerMacs whatever. The sad situation now is that in terms of compute power, Apple's top-end processor is approximately comparable to the bottom-end Durons and Celerons which are being sold respectively by AMD and Intel. (Except when running cooked benchmarks which only measure very carefully-selected Photoshop filters.)

And the reason why is that Motorola can't get the faster ones to work, and deliver them in quantity. Apple's been fucked by this for three years now, and it's one of the things which has contributed highly to declining market share.

And it's not likely to get better. Modern top-end processors are grossly expensive to design and get into manufacturing, and Motorola's PPC business has collapsed. Apple is their single biggest customer but still makes up a very small portion of the whole. To make it financially successful for Motorola, there needs to be a certain large volume and most of that was coming from customers using the PPC in embedded applications. But the PPC there is largely being replaced by ARM, and Apple simply doesn't do enough volume by itself to support a separate CPU architecture.

Steve Jobs is dancing as fast as he can; he's doing what he can to mitigate the damage. But in the long run the only solution, besides praying that Motorola suddenly passes a miracle, is to aba