I'd eaten barbecue with some microprocessor guys from Motorola. Bill Davidow's book, Marketing High Technology, had come out the year before, and we'd talked a lot about it. Everyone in the office got their own copy, they said. They were still hot.

Not surprising. Davidow, a former Intel marketer, documents Crush, the focused campaign that landed the 8086 inside the original IBM PC-at the expense of Motorola's 68000.

Though the book is still a must-read, a key piece of the winning formula-designing your own microprocessor and lining up a dozen alternate sources to secure the big contract-sounds so 20th-century in these IP-everywhere days.

Fifteen years later, the PC market that Motorola and Intel coveted is maturing and The Next Big Thing has not come into focus. But the battle lines for control of its innards are already drawn, with Intel and Motorola in the thick of it. As it happens, they are the only two companies in a sea of ARM licensees that hold rights to tinker with the

architecture.

At this juncture, Motorola has decided to leave the architecture alone. In that sense, it is following Davidow's advice to win customers with a stable of alternate sources. That route may be good for ARM. But Motorola is undercutting its own prospects for world domination by enabling any and all ARM licensees to compete with it for business.

Intel, on the other hand, is altering the architecture-and in doing so, adding a proprietary element. It has lengthened the architecture's pipeline and imported multimedia extensions from its PC processor lineup. (TI is going down a similar road by extending ARM capabilities with its DSP technology.)

Intel is clearly hoping that the ARM foundation for its Xscale processors will give the appearance of an open architecture. If it can garner enough early design activity-such as the wins it has collected in handhelds-then it can cajole operating-system and applications vendors to optimize for its seven-stage pipeline and to use its multimedia extensions. That will give Intel a clear advantage over its competitors.

Is that how it will turn out? Or, will a pure ARM implementation prevail? We'll see. In a battle like this, market momentum will determine the winner. That's why design wins are so crucial this year, as the market congeals.

This much is certain: Loser buys lunch.

Mike Feibus is Principal Analyst at Techknowledge Strategies Inc., a market research firm in Scottsdale, Ariz., that focuses on components for mobile systems (mike@techknowledge-group.com).