They Called Him Mr. USB

     The head of the USB initiative at Intel was Jim Pappas (below), a gentle giant at six-foot-seven who seemed to always have a joke or anecdote handy.

Jim was a rare combination, a chip engineer and a people person. He had spent four years designing at Digital Equipment Corp. (DEC) a single chip processor for its popular (but quickly becoming passe) Vax minicomputers. His good work won him an offer to design one of the new Alpha CPUs DEC was working on to rival Intel’s x86.

But Jim didn’t want to spend another four years with his head buried in the details of processor microarchitecture; he wanted to work with people. So, he took on another job: rallying DEC’s engineers around a new bus, a digital channel for linking processors. It would replace four different buses DEC had designed solely for its own systems.

Early on in the work, “someone on my team met with Intel and said they had a new bus that could revolutionize what we were doing at DEC,” Jim told me over coffee recently.

He read the engineering specs on what Intel was calling LGB for Local Glueless Bus. “I said the tech looks good, but the name sucks,” he quipped.

In late 1991, it got a new name, a shade less nerdy -- the Peripheral Component Interconnect (PCI). It was the kind of simple, descriptive handle engineers liked.

PCI was the first major initiative from Intel Architecture Labs, a division CEO Andy Grove created. Like its name suggested, Grove directed the group to be the architects of the PC for the entire industry. He wanted to drive the technology forward to expand the market. Over time, his decision proved to be a two-edged sword: PCs advanced rapidly but Intel’s customers, the PC makers, struggled to differentiate systems that Intel was reshaping at a heady pace.

Initially, the new technologies were well received. Pappas, for example, saw the value of plugging DEC’s systems into PCI, a bus that an emerging industry would build products around. He became one of the first and most vocal members of the group formed to promote and advance the bus. “I met with nearly every company in the industry to make sure they adopted PCI,” he said.

His advocacy attracted Carl Stork’s attention. They met at one of the early WinHEC events.

“Stork asked if the PCI bus would be open or would Intel charge royalties. I said it was completely open,” Jim recalled. “Stork wanted to talk to me because DEC said we’re putting it on our Alpha and Vax computers, and that was a big validation point encouraging Microsoft to get involved.”

Eventually chip, board, systems and software makers -- the whole PC industry -- got behind PCI. The director of the Intel Architecture Lab called Pappas at DEC to thank him for his efforts recruiting others and to offer him a job driving USB, his next big initiative, a cousin to the internal PCI bus that would fix the mess outside the box.

Pappas jumped at the chance. His enthusiasm for the job led to him being remembered to this day as Mr. USB.

Though people generally liked the idea of USB, the road to adoption was bumpy. PC users wouldn’t buy new systems unless they could plug their old keyboards, mice and other gadgets into them, too. So, system makers supported the older ports for a long time, slowing the transition to the new bus.

The good news was USB enabled some cool new stuff. “When we started working on USB, there weren't digital cameras and MP3 players. What we did helped enable these systems,” Jim told me for a 2004 story.

A small sampler of USB devices

The new gadgets wanted even faster speeds, forcing Jim’s team to quickly ramp up work on a version 2.0 spec. Opening up a popular spec to changes can be an invitation to chaos. But Jim could be firm when he needed to be. For example, some folks wanted an industrial version of USB that could carry signals a kilometer. No way, Jim told them.

The new version, and its follow-ons, were hits. They enabled new things like thumb drives aka flash drives, devices barely larger than the USB connector (or a thumb) that let users carry gigabytes of stuff in their pockets. Flash drives married two innovations: USB and the flash memories invented by folks like Eli Harari, the guy I had interviewed at my first Comdex. The mini-drives became so cheap and popular companies would hand them out as freebies at events.

Today billions of USB devices ship every year for all sorts of sometimes strange uses from sex toys to smartphone chargers.

PCI and USB were the first of many standards in a laundry list of improvements to the PC kickstarted by engineers at Intel and Microsoft, often inside Intel Architecture Labs. They set the formula: Wintel engineers would determine what the community needed, write perhaps eighty percent of the specifications for it, then pass it on to Jim’s and Carl’s teams to get PC makers to adopt the technology.

In most cases, including USB, the work served the common good. Early PCs used a rat’s nest of different cables for keyboards, monitors, printers and other add-ons. Some used serial, parallel or other ports; Apple had its own legacy and upcoming new connections. USB standardize all that into one link everyone could use, driving down costs and eliminating user confusion. It was a win-win-win.

Next: Lumbering toward multimedia