Intel’s 40th anniversary is July 18 and the world’s largest chipmaker is making a big bet on the past.

The Santa Clara, Calif., semiconductor maker would undoubtedly quibble with this characterization of its strategy for the next four decades. However, Patrick P. Gelsinger, an Intel senior vice president and one of the company’s leading technologists, stressed the crucial importance of the backward compatibility of the company’s future microprocessor chips during an informal press briefing he gave Monday to commemorate the company’s upcoming celebration.

The idea, according to Mr. Gelsinger, is that one of Intel’s chief advantages is that Microsoft’s MS-DOS operating system will be booting on the teraflop Intel microprocessor in the personal computer of the future.

Mr. Gelsinger has been at Intel for 29 years, and it was informative to hear him describe how he came to adopt the company’s compatibility orthodoxy while he was working on the design of the 386 microprocessor during the 1980s.

Faced with the prospect of having to create a circuit to do EBCDIC [an IBM data-encoding standard from the 1960s] adjusted arithmetic, Mr. Gelsinger took some pains to convince an Intel marketing executive that it was foolish to maintain the ability to support that mode. Most of the EBCIDIC programmers would soon be dead, he suggested.

“I went and pounded on our marketing manager at the time,” he said. “Why on earth don’t we drop this stupid compatibility mode?”

In response, his manager marched into his office and wrote a single word on his white board: COMPATIBILITY. The two continued to debate the point for a month before Mr. Gelsinger was ready to concede the value of supporting all the software of the past.

Now, however, he is a believer. “At the end of the day, we have no idea when the last EBCDIC adjusted arithmetic is required,” he said.

To be fair to Mr. Gelsinger, backward compatibility was only one of the four predictions he made about the next four decades on the chipmaker’s road map. The other three include the continuance of Moore’s Law with all of its economic implications; the shift to many core microprocessor designs to insure that computing performance keeps up with Moore’s Law; and finally ubiquitous computing, which he basically described as Intel chips everywhere.

Ubiquity will be the company’s biggest test. Intel is busy trying to drive the power usage of its chips down dramatically enough to make it possible to enter new markets beyond personal computers and servers. Mr. Gelsinger and other Intel executives argue that the vast libraries of Intel-compatible software that run on his chips will be one of the company’s key competitive advantages.

However it is already clear that it won’t help the chipmaker win in every competitive situation for the new markets it wants to win.

Mr. Gelsinger acknowledged that Intel had lost the heated bidding for the next generation of Apple’s iPods and iPhones.

“Apple chose not to take that roadmap at their next generation of that platform,” he said, adding: “I don’t think that means that we’re not ever going to have them on that roadmap.”

Copyright 2008 The New York Times Company