July 6, 1999
Web posted at: 2:47 p.m. EDT (1847 GMT)

by Leslie Goff

(IDG) -- Neither Bob Grimm nor Roy Clay nor any of the other folks on the team that engineered Hewlett-Packard Co.'s first foray into the computer market in 1966 would have believed it if someone had said that HP's instruments business would one day be computing's weaker sibling.

"I don't think anyone at that time dreamed that the computer business would take over. Not even Bill Hewlett or Dave Packard," says Bob Grimm, who in 1966 was general manager of HP's Dymec division. "They were very farsighted. But the idea that we would split the company some day and the HP name would go with the computer business would have been laughed at."

Debuted in November 1966, the 2116A was designed to automate the collection and processing of data from the company's test and measurement devices. Without realizing it, architect Kay Magleby and the team that developed the 2116A ultimately altered HP's future when they decided to build a general-purpose computer that could be tied to myriad instruments via plug-in cards rather than creating a very device-specific machine.

The target customers for the 2116A were HP's bread-and-butter engineers -- the folks who had to measure, test and calibrate. They could hook HP's nuclear scalars, electronic thermometers, digital voltmeters, pressure transducers -- anything that collected data -- to the computer using the cards, which featured only a cable, the connectors and the hardware interface. The data collected could then be manipulated and compared with data collected earlier, via the software; the results were printed to magnetic or punched tape or to a Teletype machine.

The plug-and-play ease with which customers could connect their existing instruments to the 2116A was a key marketing point -- and also the feature that changed HP's destiny.

"The idea of having a small computer owned by a department was a radical concept," Clay explains. "And in a lot of companies, if you wanted to buy a computer, you had to go through a hierarchy. You couldn't just sell it to the person with the problem like the instrumentation products."

And it was a hard sell internally, too. HP sales managers "were not at all excited about it. They didn't want to use any of their best field engineers on this new gadget," Grimm says. He managed to recruit four people, whom he put through six months of training on the 2116A. They backed up the sales representatives and drove the 30-in.-high, 230-lb. computers around in station wagons to customer sites for demonstrations.

The 2116A project was "radically different" from other areas of HP. Its staff saw the potential applications of the 2116A as more than just a back-end for instrumentation, says Clay, who led software development and is now president of ROD-L Electronics Inc. in Menlo Park, Calif.

Even though HP had no intention of selling the machine as a general-purpose computer, it was purchased by several companies, including Tymshare, as the basis for time-sharing systems. So when Holiday Inn came to HP seeking a fail-safe computer for a point-of-sale system, Clay's team tied two boxes together in a parallel-processing model. If one failed, the other would take over. Holiday Inn was sold on the idea, but HP's senior managers were not; the system had nothing to do with the original purpose of the 2116A.

"Bill Hewlett called me," Clay recalls, "and he said, 'Cancel the order and terminate the project.' And at 3 p.m., it was terminated." HP wouldn't commit to the computer industry on a significant scale until 1975, when it introduced its HP-3000 series. More than 20 years later, in 1999, HP would reorganize to formally acknowledge the computer business as its core product line.

"The recent split is at least 180 degrees different from 1966," Clay says.

Goff is a frequent contributor to Computerworld. Contact her at lgoff@ix.netcom.com.