The last time I wrote about a keynote by Apple's charismatic chief executive Steve Jobs, I made one foolish mistake. I wrote my piece before Jobs had spoken.

I predicted - no, perhaps not as foolish as that - I "assumed" that Jobs would be encouraging independent software authors to switch from writing programs for Windows to writing for his company's platform, Mac OS X. I thought he'd exploit the long history Microsoft has of undermining third-party developers by studying their products, then incorporating them wholesale into the next Microsoft Windows release.

Jobs's keynotes are famous for their surprises. As it turned out, he did the exact opposite of what I'd imagined. He announced a raft of free or cheap Apple software products, all of which extinguished the market for various small independent developers.

In one flagrant case, he announced an extension to MacOS X that cloned almost exactly a product called Watson, a $30 (€26) add-on marketed by Karelia, an independent developer to whom Apple had previously given a design award.

Karelia watched that last keynote, I imagine, aghast as its potential audience evaporated before its eyes.

Such death-by-flattery tricks do not appear to have diminished the programming community's love of the Mac - or of Jobs's keynotes. In the past year, the number of applications for the platform has doubled to 6,000.

Last week, I joined hundreds of the coders of those apps, milling in downtown San Francisco for Apple's Worldwide Developers Conference. They had each paid over $1,000 for the privilege of attending and hearing Apple's technical staff give them advice on how to program for Apple machines. And, of course, the chance to hear Steve Jobs deliver his keynote.

This time, I was smart, and I have saved my clever-clever analysis until after Steve spoke.

As it is, I need not have bothered. The biggest secret of the string of Apple announcements Jobs revealed had already been leaked a week before. A Web administrator at apple.com copied the wrong image to the wrong server, and accidentally revealed this year's coup de main: a new processor for Apple to compete with Intel and a new top-line desktop PC to compete with the most powerful Windows machines.

Expected or not, both of these are extraordinary advances for Apple. As recently as a year ago, many analysts were sceptical that Apple's desktop computers could ever catch up with Intel-powered machines in computer speed.

Intel produces the x86 chipset, used in every Microsoft-powered machine since 1981. It is Intel's core business: its chips are in nearly 150 million machines every year. Intel throws nearly $4 billion into x86 R&D. It plans three generations of x86 processors in advance.

Apple, on the other hand, has been dependent for its processors, the PowerPC series, on two external companies: Motorola and IBM. Neither of them are dependent on the PowerPC for their profits and Motorola, at least, had appeared to have grown weary of investing in a chip that is almost exclusively used by Jobs's tiny segment of the personal computer market (there are five million Mac OS X users, total, in the world).

Instead, Motorola concentrated its PowerPC development on requests from larger clients who use the chip in embedded applications - dedicated industrial equipment, mainly - cars and aeroplanes, in other words. These customers are not interested in speed. They want processors that use a low amount of electricity.

The upshot of this has been that, while Apple's laptop market has been able to take advantage of Motorola's experiments in low-power processing, its supposedly super-fast desktop machines languished behind Intel.

Jobs's announcement changes that. The new top-of-the-line Apple machine is based on new IBM processors. IBM's primary interest in the CPU arena is to power its high-powered scientific workstations. Put baldly, the new Apple G5 machine is now dependent not on Motorola's souped-up car chips, but on IBM's cut-down supercomputer chips.

At any other event, this would be the highlight: the big launch. But Jobs's detailed run-through of the technical merits of the processor provoked a rather muted response from his audience. Partly, I'm sure, this was due to them all knowing in advance thanks to that careless - and now, I imagine, buried under the Golden Gate Bridge - webmaster.

But principally, it was because these were software developers. They do like their fast machines, but many of them would never be able to afford even the cheaper G5 machines.

Instead, the biggest whoops came when Jobs demonstrated the latest eye-candy written by Apple's own programmers. If there's one thing developers for his platform have, it's inspiration from the front.

Jobs seamlessly ran through seemingly trivial new operating system features - such as a method for choosing between windows on a crowded desktop, or quickly switching between users. Both of these are new to the Mac, but rather old hat to Linux or Windows users.

But Apple's implementation has style: its window-switching involves a cinematic pull-away from the screen effect; its user-switching flips the desktop around like an election night TV graphic. They look snazzy, and they work well.

It's aspiring to emulate that combination of utility and polish that has many of the coders attending these conferences. And they seem particulary cheered when Jobs revealed similar new sleekness to the programs that coders have to use to write applications on the Mac.

The new development system will allow coders to dump the slow process of compiling a new program onto other people's idle machines on their network - in much the same way as film studios use large banks of computers to render individual frames of computer graphics.

The new environment, called XCode, also allows them to fiddle and fix bugs in a program as it is running. This contrasts with the traditional method, whereby bugged programs have to be halted, taken apart and then reconstructed every time a problem is found.

It's attention like this that makes me think that perhaps I was right about Jobs after all: he does realise that he needs to attract developers to make his machines a success. And that, amidst all that show biz, he has staff working hard to pull them into the Mac world.

That's important, because without fresh new applications, the fastest processor in the world isn't going to help Mac steal more of the market.