Last year, while working as an (eventually unpaid) analyst, I told some Intel executives they needed to embrace a "platform" strategy for their product.

My call was ignored. The junior executives I was writing for could afford to do that. I was just an analyst.

Fortunately there was a larger bullhorn, that of incoming CEO Paul Otellini (left, from, UC Berkeley) and now Intel is embarking on precisely the strategy I called for.

What does it mean?

Intel doesn't sell chips that make products. It makes platforms on which generations of products can be based.

Take the ixp425, Intel's failed communications processor. It's a nice chip. In many ways it's structurally sounder than the offerings of its competitors. But it went nowhere because competitors would deliver complete reference designs -- hardware, software, schematics. Intel told the channel to do this itself.

A platform strategy is a form of jiu jitsu on this problem. By developing, not just a chip but a platform, on which many successive generations of products can be built, Intel gives its customers the ability to provide long-term value. What is sold as, say, a home gateway can quickly be upgraded into an Always-On interface.

What does it take to build such a platform? It takes a chip capable of doing many different things, and Intel's chips have that unique capability. But it also takes a robust, scalable, modular operating system, like Linux, on which today's applications can be written, and on which tomorrow's applications can be built.

A product based on a reference design may hit the market faster. It may cost less. But if you have to replace it next year, and the next year, and the year after that, there's no value.

On the other hand, if you buy something baed on a platform, you can build on it. What you buy next year and the year after can work with what you have now, seamlessly.

You only get that with a platform.

The PC was a platform for the 70s.

The network was a platform for the 80s, built on the PC platform.

The Internet was a platform for the 90s, built on the network platform.

What Intel is committing to is platforms built on all those other platforms.

And that's the difference between a computer, something that lasts, and a consumer electronics product you toss away.

Note: I know Otellini is not a Truly Handsome Man (like I am), but if he can overcome the handicap of hair with raw brainpower, who am I to begrudge him his unfortunate locks, uh, looks?