Processor pioneer Victor Poor dies of cancer at 79
Modest designer of Intel's first integrated processors
Obituary Victor Poor, whose death was announced this weekend, isn’t one of the first names to come to mind among computing history, but he played a pivotal role in the development of Intel's early architectures that went on to dominate modern computing and is a legend in ham radio circles.
Poor, along with fellow radio enthusiast and student Harry Pyle, designed what became Intel's 4004 chip, on Thanksgiving weekend in 1969 – sketching out the architecture in his living room in a marathon four-day session. He also played a key role in working with Intel on the 8008 chip, the world’s first 8-bit microprocessor.
Born in 1933, in the midst of the last Great Depression, Poor showed an early interest in technology, describing himself in a 2004 interview with the Computer History Unit's oral history division as a "natural-born nerd" who lived for ham radio.
After a stint in the US Navy following World War II he qualified for naval electronic school before working on the first Univac systems and the Packard Bell 250 computer. He recalled that the Univacs, with their delicate construction of vacuum tubes and liquid mercury memory, were prone to breaking down every 90 minutes or so, so any programming had to be quick and correct. The team was working on a computer program for air traffic control at the time, which caused him some concern for the safety of passengers coming in to land.
The same reliability problems of analogue computing dogged his next job as a field engineer for Raytheon, fixing up the Sparrow III air-to-air missile. The missile's control systems relied on the analogue circuitry and vacuum tubes and "didn't work very well," he recalls, particularly in tropical parts of the world.
In 1959 he left to start up what became Fredrick Electronics with $100,000 in VC funding. Much in the same way as Hewlett and Packard, Frederick was a radio and telecommunications engineering shop, working on military and commercial communications and building some of the first reliable telex systems capable of transmitting images and sound.
Occasionally these products were accidental. Inspired by his ham radio systems, Poor built a machine to convert teletype into Morse code. This was just something he was noodling with, but a visiting US Navy officer saw it and insisted on ordering some, and eventually the company sold huge numbers of machines that could convert Morse code into teletype and vice versa, eventually achieving transmission speeds of 300 words per minute.
But these devices were still analog and weren't programmable, running purely off ROM memory. This led to significant reliability problems. While machine to machine communications were reasonably reliable, the tapping style (or fist) of a human Morse code operator could completely flummox the machines.
After the company was bought out by Plantronics, Poor turned his mind to integrated circuitry, and joined Computer Terminal Corporation (CTC), a startup by two NASA engineers. The company pitched Intel and Texas Instruments to build the 4004 processor design he and Pyles had developed and used it in a range of teletype machines based around integrated circuitry.
Poor also worked with Intel on the next generation of processors, which became the 8008 line that helped spur Chipzilla's massive growth. CTC again asked Intel and TI to try and build the chip, eventually using most of Intel's offering but with its own tweaks. Intel bought the rights to the hardware in exchange for writing off its original $50,000 fee.
Despite his pivotal role in the design of early computer hardware Poor was very modest about his input, typically shifting the praise to fellow engineers who also worked on the project. Patenting such systems wasn't also something he had much time for and he appeared to regret the current mania for owning every new technology in a legally binding way.
CTC was renamed Datapoint and developed its own processing systems under Poor's guidance. But the company decided not to follow IBM's move into the personal computer market, for fear that it would undercut its own highly- profitable business – a decision Poor personally disagreed with.
He retired in 1984, saying the company had got too big and he wasn't having fun anymore. For someone with more experience than most in the field, he described himself as still clueless about how to run a technology company efficiently, saying he felt too much like Dilbert's pointy-haired manager.
"I've often thought about what's the good way to organize a company and I've come to the conclusion there is no good way," he said. "No matter what we did, you know, at Datapoint we tried all kinds of schemes, we did skunkworks projects and we had separate departments and we merged departments and we unmerged departments and no matter what we did there was always something wrong."
He professed irritation with management guru Tom Peters, who interviewed him and other Datapoint management for the book In Search of Excellence. Peters misrepresented the management process, he said, making it out to be easy.
"I didn’t even recognize myself in there. He made it look like we were geniuses and it just is not true. It’s just not true," he said.
After stepping down from Datapoint, Poor and his family spent much of their time sailing in the US, Mexico and Europe. While on board he developed a communication system that married amateur radio stations transmissions and messaging servers to allow a basic email, graphics, and data transmission, which is still used by over 100,000 hams today.
In May, Poor was diagnosed with stage four pancreatic cancer and he died on Friday in Florida. He is survived by Florence, his wife of 60 years, and their two children. ®
Fascinating history, and I never knew any of it!
...and Hans Camenzind died the other day. He invented the 555 timer IC.
How come the bio in the NY Times is so different?
I'll take a wild leap of faith and say TheReg's story is a lot closer to the truth, but still amazing how they seem to be 2 very different stories.
So the 4004 and 8008 were Poor processors. The 8086 was just a poor processor.
Yet another innovating tech pioneer...
passing over with little or no marking by big media.