Catching Up With . . . David Spencer
Ever Arrive Safely from a Plane Trip? Thank Dave. Seriously.
by Brooke C. Stoddard '69
Dave was born in Pittsburgh and grew up on the grounds of the Tuberculosis Hospital of Pittsburgh in the Hill District. His father worked there as a physician and the family – Dave has a younger brother – had an apartment on the grounds. But tuberculosis was succumbing to antibiotics, so Dave's father joined the U. S. Navy for two years, practicing at its St. Albans Naval Hospital on Long Island, with the family living in Levittown. Dave's father then moved to the medical department of New England Telephone, where he retired many years later as Medical Director. This resulted in a move to Bedford, Massachusetts, outside Boston.
Both Dave and his brother were fond of science and electronics, read science fiction, and received good educations at Bedford High School. By this time, Dave was academically oriented to physics, astronomy, and electronics. He was accepted at MIT but understood about himself – perhaps owing to a high school English teacher who versed her students in challenging ideas and good drama at Boston's theaters – that there was also some interest in humanities knocking around inside. He visited Princeton in the spring, when the wisteria would enchant anyone, and he knew Einstein had lived and worked in Princeton, so he accepted at Old Nassau. Most of his years at Princeton he had a single, but sophomore year he shared a suite with Ken Braly, Bob Bridges, Bob Fletcher, Bob Meushaw, and Dave Sparrow.
While very interested in the theoretical and philosophical aspects of physics, Dave suspected the difficulty of his making a living in that field, so he opted for a double major – Electrical Engineering and Physics. By Junior Year, however, he found this workload was oppressive, so he dropped the Physics segment.
He also became interested in computers. At the time there was no computer science department and in fact there was debate over whether computers should be the subject of an academic department (it became so at Princeton in 1985). "Of course, I was wondering 'What would I do next?' when I discovered a book called Computers and Thought by Edward Feigenbaum and Julian Feldman," Dave recalls. "It was about artificial intelligence and became a classic.” Dave was still exploring his other interests, and in the summer of 1968 interned at the National Radio Astronomy Observatory in Charlottesville, Virginia, rooming with classmates and fellow interns Ken Braly and Jay Gallagher. “That was a memorable summer for me, not only for the national events taking place, but also because I got to see the large radio telescopes in Green Bank, West Virginia, up close, and to ‘play’ with the computers that controlled them.”
There was also a tutorial program, an introduction to radio astronomy by leading researchers. Just the year before, pulsars had been discovered, and it was still not clear, when the work was presented to the interns, whether these rapid periodic radio pulses from out in space were from a natural phenomenon or from “Little Green Men”—in fact, the first discovered pulsar was for a while called LGM-1. Earlier that year the movie 2001: A Space Odyssey, with its intelligent HAL 9000 computer, came out. “I saw it that summer with some of the other interns," Dave recalls. "Seeing it in that context made it even more exciting. By the time I returned to Princeton I had decided to focus on computers and AI.”
At the time, no AI work per se was done at Princeton, so Dave's senior project – for Professor Theodosios Pavlidis – was attempting to determine if computers could equal seasoned engine mechanics at determining an engine's problem by listening to the sounds it made. His summer job after graduation – for Professor William Surber, Jr. – was to add circuit drawing capability to an existing computer simulation of electrical circuits using one of the early IBM interactive computer graphics displays. He was working on that when he and friends from the Princeton computer center drove to Cape Canaveral, slept on the beach and witnessed the launch of Apollo 11. Days later, he watched the moon landing on TV in one of the Butler Apartments graduate student housing units, his summer living quarters. It was another memorable summer.
Dave was admitted to graduate work at both MIT and Carnegie Mellon and selected MIT. He shared an apartment with classmates Jeff Kaplan and David Pensak. After about a year, it became clear that a PhD was not for him. It seemed that all the exciting things people imagined for AI, like intelligent mobile robots, autonomous vehicles, and the HAL 9000, were not possible given the technology of the time. What was possible at that time did not seem very interesting to him, and Dave had no idea how to bridge that gap. As it turned out, many of those AI dreams started to become possible only as Dave was considering retirement. Rather than pursue a PhD, Dave felt the call of making a living and building things that were more immediately useful. He left MIT with a Master's Degree plus a few extra courses that got him a degree unique to MIT and one or two other universities called Electrical Engineer.
As it turned out, MIT's Lincoln Laboratory, which normally did work for the Air Force, was just then branching out to start a major project for the Federal Aviation Administration. "Because I had just learned a good deal about computer graphics and working with real-time computer systems, I was a natural fit for this project," Dave recalls. "Lincoln Laboratory accepted my application and I stayed there my entire career."
Radar-based air traffic control can be traced back to the Battle of Britain. The Royal Air Force in 1940 had a radar system that could detect aircraft within the radar’s beam out to a certain range but could not distinguish friendly aircraft from foe. So, the RAF placed in their aircraft devices called transponders that, when stimulated by a special signal from the radar, sent a distinct signal back; the signal identified them as RAF planes and not Axis ones. This Identification Friend or Foe (IFF) technology survives in modified form in civilian air traffic control as a system called the Air Traffic Control Radar Beacon System (ATCRBS). ATCRBS ground interrogators can read out an aircraft’s identity code and its altitude. A problem is that within the reach of an interrogator’s antenna beam, multiple aircraft transponders send back overlapping signals in response. The more aircraft that are within the antenna beam, the more jumbled the returning signals. The challenge of Dave's FAA project – initially called the Discrete Address Beacon System (DABS), now deployed worldwide as Mode S – was to create the capability of addressing interrogations to aircraft individually, allowing the ground station to schedule interrogations so that the replies didn’t overlap, and providing a digital data channel in both directions between the ground station and the aircraft. Dave was the lead software engineer for the experimental prototype DABS ground interrogator that Lincoln Laboratory built for demonstration to the FAA. Later he helped transition the technology to industry.
"The next step after that was to build a system such that each airplane could have its own miniature short-range Mode S interrogator," Dave recalls, "to identify the location and altitude of airplanes nearby, exchange data with similarly equipped aircraft, provide a traffic display to pilots, and, if necessary, issue climb or descend advisories to the pilots, thus providing a collision-avoidance system." This became TCAS (for Traffic-alert and Collision Avoidance System), internationally known as ACAS (Airborne Collision Avoidance System), now carried on almost all commercial aircraft. It operates independently of the ground-based air traffic control system and air traffic controllers. Dave was again the lead software engineer for the development of the experimental prototype TCAS units and played a role in flight experiments and data analysis.
After the success of TCAS, Dave worked on a couple of other experimental air traffic control related systems, but these ultimately had little success. He got a chance to jump back into the AI field for a while in the 1980s, when there was an explosion of interest in practical applications of AI. Dave and his team did have some success at automating the job of an air traffic controller using the “rule-based expert system” technology of the time. However, that hand-crafted approach wouldn’t scale to the entire air traffic control system, and it was unclear how to assure safe outcomes in all situations. Other researchers had similar results. As a result, to this day, humans are still the heart of air traffic control. Similar problems in other domains led to the collapse of the AI boom altogether.
Dave’s work eventually changed direction from designing software to analyzing failed software projects for the FAA, trying to determine whether anything could be saved. That led to some very interesting work related to the Cassini spacecraft – launched toward Saturn in 1997—an enjoyable assignment that took Dave to the Jet Propulsion Laboratory. "I'd always been interested in spacecraft, and would have liked to work in that area," he says, “but there did not seem to be any openings at the time I graduated.” Cassini was a very large and expensive program to build a spacecraft that would have to operate without major failures for many years. NASA was determined to avoid the Hubble Space Telescope mistake of launching a flawed product and so commissioned a review by outside experts. “Our job was to make sure the people at the Jet Propulsion Laboratory knew more about their spacecraft than we did,” jokes Dave. Whatever Dave's contribution, Cassini was a remarkable spacecraft; it took astonishing images of the Jupiter and Saturn systems and lasted 20 years, only recently plunging into Saturn's clouds.
In 2004, Dave was ready to move to a lighter schedule. Fortuitously, just at that time the perfect semi-retirement job appeared. The European air traffic control organization EUROCONTROL came to the FAA with a proposal to amend the TCAS decision-making. The problem was that TCAS would not “change its mind” and reverse its recommended climb-descend advisory if the pilot, or the pilot of another TCAS-equipped aircraft in a potential collision situation, failed to follow the advisory or moved opposite to the advisory. This flaw was a secondary factor contributing to the collision of a cargo jet with a large passenger jet near the town of Überlingen in Germany in 2002 with the loss of 71 lives. The FAA quickly assembled some of the original TCAS engineers like Dave who were in or near retirement and coupled them with younger engineers to craft the necessary adjustments. This led to a desire for other improvements, and eventually to a proposal for an entirely new design for the collision avoidance algorithms called ACAS-X. However, Dave was interested in full retirement by then, and knew that getting involved in ACAS-X would delay that, so he instead cheered on the younger engineers from the sidelines and stuck to answering questions about the original TCAS.
Eventually, those TCAS questions faded away and full retirement got him in 2014. Recently he and his wife Beverly quit Boston for Silver Spring, Maryland, to be close to their daughter and her family, including two young grandkids. Helping with child care is now a big part of their lives. Genealogy and family history take up a good bit of Dave's remaining free time. The growing volume of online historical and genealogical documentation allows him to do a great deal from home. However, not all documents of interest are online, and he and his wife have also made research trips to dusty archives, local libraries, historical societies, deed registry offices, and cemeteries looking for clues. They have also jumped into genetic genealogy, taking DNA tests provided by Ancestry and other companies, and have obtained interesting and useful results. Among other things, Dave finds he has (very) distant Australian cousins. And no, it appears those cousins’ forebears went there voluntarily and not on convict ships. He also follows current theoretical and experimental physics, as well as other sciences, at the level of Scientific American magazine and books not only by Stephen Hawking, but also by Brian Greene and by Lisa Randall on string theory. Also, he occasionally flies; he has confidence in the technology.