NASA Office of Logic Design

NASA Office of Logic Design

A scientific study of the problems of digital engineering for space flight systems,
with a view to their practical solution.

Profile Apollo: Hugh Blair-Smith

Elwin C. Ong, MIT
April 10, 2005

“One small step for man, one giant leap for mankind.” Nearly four decades after those words were uttered by Neil Armstrong, many people around the world have equated that singular statement as human kinds' ultimate achievement in the 20th century. So, how was the greatest feat in the 20th century accomplished? Opinions vary, but the most popular consensus points to the political atmosphere of the 1950's and early 60's. The cold war and technological competition between the United States and Soviet Union provided the prime motivation for Apollo. It was a competition that held severe consequences for the future of political thought. The country that won the space race would, naturally, win the ideological competition for the superior form of political governance. While political motivations to win the space race cannot be dismissed as a prime factor for Apollo, it is arguably less significant than what most political historians have attributed to it. Among the people most familiar with the Apollo project, those who had their hands in the design and engineering of Apollo, one often finds a more considerable motivation.

“It was for Jules Verne and Kepler and Goddard.” [HBS] Hugh Blair-Smith, an engineer who played a major role in the development of the Apollo Guidance Computer will tell you that the race with the Russians was secondary at best when it came to the motivations for sending men to the moon. The foremost reason was one intrinsic to every engineer, mathematician, and scientist. The mid-twentieth century was the first time in human history where the dream of reaching the moon, as people like Verne, Kepler, and Goddard had imagined it became technologically, and humanly, feasible. Much as compasses, celestial navigation, and advancements in ship-building technology allowed Columbus to sail across the Atlantic in the fifteenth century, inertial guidance and navigation, material sciences, and advancements in computing finally allowed engineers to realize the dream of sailing across space. “Nobody had done what we did, but we knew we could do it.” For many Apollo engineers like Hugh Blair-Smith, this statement came to represent their primary motivation throughout the program.

Ever since he was very young, Hugh Blair-Smith always wanted to be an engineer. As a kid, he had imagined himself building bridges crossing Eastern Long Island Sound. Little did he know that his engineering career would be significantly different than the one he'd imagined. After moving around several cities along the East Coast including Washington and Philadelphia, Hugh began his Massachusetts residency when he began studies at the Middlesex School in Concord. After graduation, he enrolled at Harvard University, graduating in 1957 with an AB' in Engineering and Applied Physics. At Harvard, he would find a permanent zip code, 02138; he prides himself on never having moved outside of Cambridge. At Harvard, he would also start his long career as a computer scientist, and within a few years, began work on the most exciting computing technology of the day, the Apollo Guidance Computer (AGC).

The state of computing technology was progressing at a steady pace in 1957, Hugh's senior year at Harvard. Just two years prior, in 1955, the world's first digital computer, ENIAC was officially retired. The designers of the ENIAC, J. Presper Eckert and John Mauchly had also produced the UNIVAC I, the world's first commercially available computer in 1951. By 1957, there were a total of 46 UNIVAC I's [THO]. One was donated to Harvard University where it was programmed by Hugh and other undergraduates in their computer science courses. The UNIVAC I and those computer courses would point Hugh to his career path. Within a year after graduation, while moonlighting on various programming jobs, he became so motivated that he went around to various computer companies and begged for programming manuals. “So I believed at one time, back, maybe up to 1958 some time, that I had some knowledge of how to program every computer that was abroad in the world.” [HRST] His motivation and hard work was rewarded when he was recruited by Dan Goldenberg to work for his computing group at MIT's Instrumentation Lab.

Hugh began his career at MIT in September of 1959, two years before the Apollo contract was assigned to the Instrumentation Lab. The major computing project at MIT at the time was the Mod 1, a computer that would have been used on a proposed robotic mission to Mars. The Mars probe had to be completely self-sufficient. There was no way to uplink commands to the spacecraft [BAT]. The Instrumentation Lab, under Charles Stark (Doc) Draper had pioneered the field of inertial guidance, and the Mars probe would have been the perfect application. Hugh was hired to write an assembler program for, as he describes it, “an unknown number of machines with unknown characteristics.” [HRST] The goal of the Mod 1 computer was to have both the hardware and software, including the assembler program ready by Christmas of 1959. Hugh fittingly named the assembler program YUL. YUL initially ran on an IBM 650 and produced punch cards that were later translated into instructions for the Mod 1 computer.

As the Mod 1 evolved into the MOD 3C, Hugh became involved in the design of the instruction sets. Along with Al Hopkins and Ray Alonso, they designed the instruction set for the Mod 3C with a total of 8 base instructions. The MOD 3C, a design with a total of 4,096 words eventually became the baseline for the Apollo Guidance Computer.

The Instrumentation Lab in 1960 was very much a research laboratory. Although the lab had contracts with various government agencies, it operated much more like a university laboratory than a government contractor. The atmosphere was very informal. Hugh remembers having a lot of freedom, working among various research groups without much impedance. “The only bureaucratic bit was getting my security clearance.” [HBS] He was singularly responsible for YUL, and throughout Apollo, he maintained the program, updating it as the design of the AGC changed. In addition to being the assembler, YUL became a version control system for the AGC as it evolved.

On August 10, 1961, NASA formally selected the MIT Instrumentation Lab to provide the primary guidance, navigation and controls for Apollo. The requirements for Apollo were similar to the Mars probe and naturally; a digital computer became the central part of the system. The digital computer was the most important technology needed to accomplish the goals of Apollo. It is difficult to imagine how an analog or mechanical device could have accomplished all the complex tasks required by Apollo and still have fit within the stringent weight, size, and power budgets. With their experience developing the Mars computer and the Polaris missile digital computer [HALL] before that, the team at the Instrumentation Lab was confident that they would be able to deliver. “We knew we were the best at what we did, and that would work.” [HBS] The tone was set by Doc Draper, who volunteered himself for the mission, so confident was he that his team would deliver that he was willing to put himself at risk to prove it. Of course, he could have just been as excited about the chance to float around in space.

The design of the AGC was primarily divided into two sections. One group led the hardware development and support software like the assembler program, while the other was responsible for mission software. Hugh prides himself on having worked on both sides of the fence. While his primary responsibility was the YUL program, he also played a secondary role in the hardware logic design. In addition, Hugh also got the opportunity to write mission software for the AGC. The routine, named R29, was to be used in case the automatic pointing of the Lunar Module's rendezvous radar failed. The task of R29 was to swing the rendezvous radar around the sky in a search pattern to find the Command Module. Again, he was independently in charge of the task.

By 1964, pressures to deliver were steadily increasing. NASA began to doubt that MIT could deliver the AGC on schedule. Various committees were formed to examine the progress of the hardware and software design. Concerns about the reliability of the hardware was brought forth and there was talk of replacing the AGC with the Saturn's LVDC computer, a triple modular redundant computer used to control the Saturn V. MIT formed a team of system and software designers to examine the deficiencies of the LVDC. Hugh was part of this team that presented their findings to NASA in Houston. “When I saw how clumsy and inefficient it would be at the nimble multi-tasking required of the AGC, I made up a nasty nickname for the LVDC: “APE,” for Adaptive Polynomial Engine, in parody of Babbage's 19th-century Difference Engine Concept.” [HH] The findings of the MIT team thoroughly convinced everyone that the LVDC was limited in its capacity and capabilities and was severely inadequate for the tasks required. Reflecting on the design of the AGC hardware, Hugh believes that the reliability of the hardware and system software like YUL was never in doubt. NASA oversight did not help nor hurt the progress of the their development. Although Hugh believes the reliability of the mission software was also never in doubt, its development was much more demanding than the team at MIT first imagined.

The term software engineering was not known in the 1960's when the Apollo software was developed. “Nobody, not even NASA knew how difficult it would be to develop such a large software project.” [HBS] MIT software designers invented the software development process along the way. Software routines were divided among various engineers. Teams of programmers led by a senior engineer, developed larger, more complex routines. Since the end result of the software development process was a rope in the rope-core memory, the engineers in charge became known as “rope-mothers.” Although NASA could not apply the systems engineering approach that had worked so well in the development of spacecraft hardware, they were able to provide some much needed direction. For one, they were correct to point out that MIT was severely short on people. MIT had underestimated the monumental workload that was required to produce high-quality software. As a result, a lot of contractors were brought in to help. The rope-mothers suddenly found themselves with more sons and daughters. NASA was also correct to point out that MIT engineers had a predisposition to say yes to any apparently good idea that anybody came up with. “NASA taught us to say no.” [HBS] Perhaps it was due to the nature of the research environment, but MIT engineers were always confident they could add more requirements to the already monumental job required of the AGC. They could hardly be blamed for their inexhaustible motivation. As experts and pioneers in their field, they knew that nobody had accomplished what they were about to accomplish, and more importantly, there was never any doubt that they could indeed accomplish their objectives.

The success of the AGC software should be attributed, ultimately, to the dedication, motivation, and confidence of the people involved. As Hugh recalls, the culture at MIT fostered a tremendous significance in attention to detail. “You knew about everything in the box because you were the one who put it there.” [HBS] Every integrated circuit, and every line of code had a reason to be there, and by the time it was delivered, it had been examined and tested until there was no doubt that it worked. “The assumption and requirement was that everything had to work every single time.” [HBS] When errors were inevitably found during tests, the culture in the lab was not one to assign blame, but instead, to find the root of the error and make sure it never happened again.

For the hundreds of people responsible for the AGC at MIT, the Apollo program was a tremendous success. Each mission was celebrated and there was a great sense of accomplishment. For many Apollo engineers including Hugh Blair-Smith, their work on the program was the most significant and exciting in their careers. Hugh remained with the Instrumentation Lab well after the end of the Apollo program. In 1981 however, convinced that NASA was no longer interested in supporting research in computer science, he left to work at a startup, developing techniques for touch-screen controls.

Of the technologies required by Apollo, digital computing has proved to be the most successful in the last four decades. Much of its success is owed to the Apollo program including the development of integrated circuits and software engineering. Computing technology has had a tremendous impact on the world. Computers today are responsible for controlling passenger aircraft and subway cars, delivering information around the world on the Internet, and it is largely responsible for the creation of this paper. Comparing the current computing technology to Apollo's, there is no doubt that there have been immeasurable improvements. Yet, as computing technology has matured, the systems that are involved have become severely more complex. Unlike Apollo, where the box was filled with everything that you personally had a hand in developing; today, it may include anywhere up to “dozens of programs written by some unknown persons in Redmond, Washington.” [HBS] In such a system, attention to detail becomes significantly more difficult to apply.

A mission to return to the moon must surely experience all of the challenges faced during Apollo, but also, perhaps more significant, the designers of the next manned moon mission must face new challenges posed by complex requirements for automation and fault tolerance. The proposed Crew Exploration Vehicle (CEV) can certainly take advantage of today's powerful computing systems, but the technological challenges are still significant. Perhaps, even more daunting are the political and motivational challenges.

Apollo was a challenge to a generation of young and enthusiastic engineers willing to sacrifice a decade of their lives to satisfy the urge to do something nobody had ever done before. It is difficult to conceive that this generation of engineers are as collectively motivated about the CEV project. While no one doubts that there are significant numbers of people willing to dedicate themselves to the task of sending humans back to the moon or on to Mars, a project like Apollo requires the collective motivation of a majority of the engineering population. If you were an engineer living during the middle of twentieth century, the most interesting and exciting job you could ask for was Apollo. The most exciting computing and engineering challenges today are not related to space and exploration. It has been replaced by other fields like digital communications, the Internet, and biotechnology.

Perhaps, in an ironic twist of fate, the overwhelming success of Apollo contributed much to the decline of manned space exploration. Because of Apollo, we know today that humans can indeed land on the moon and return safely to Earth. And besides a few minor setbacks, we know this monumental challenge was conquered thoroughly. As a result, sending explorers back to the moon is no longer a novel human and engineering endeavor. The engineering motivation that was so necessary for the success of Apollo no longer exists for the CEV. As Hugh poignantly concludes, “We should not go back to the moon simply for the sake of being there when the Chinese arrive.” [HBS]

Much has been conjectured on the future political competition between the United States and China. Some believe this competition will provide the necessary motivation for the completion of the CEV and subsequent missions to the moon and Mars. The hypothesis is simplistic at best. A challenge like sending humans to Mars requires more than political motivation itself. To succeed, novel engineering challenges must be put forth and framed in the context of space exploration. Much as digital computing provided the novelty and exciting possibilities applied in the form of a manned lunar landing, today's most challenging engineering questions must be applied to space exploration. Only then will we inspire a new generation of Apollo engineers like Hugh Blair-Smith.



Home - NASA Office of Logic Design
Last Revised: February 03, 2010
Web Grunt: Richard Katz