Bill Schweber, a former magazine editor, suggested I read the book, "Digital Apollo: Human and Machine in Spaceflight," by David A. Mindell.  The book details the development of electronic controls used in the Mercury, Gemini, and Apollo space-flight programs, with particular emphasis on the Apollo program.  Younger engineers might not remember the "space race" with the Soviet Union that created a vast and intense engineering effort to land a man on the Moon before the Soviets, but it was a palpable part of the "Cold War."  (See Book Note below.)

The first part of the book details the friction between test pilots--then the professionals closest in job descriptions to astronauts--who wanted to "fly" missiles and spacecraft, and the engineers who thought of astronauts as simply passengers on automated flights.

I most enjoyed the latter portion of the book that describes the development of the small computers--actually microcontrollers--and the relatively short programs needed for navigation and guidance.  Although this work occurred over 40 years ago, we seem to go through many of the same trials and tribulations today.  And, younger engineers might think things like real-time operating systems problems and documentation snafus have short histories.  Not so.  Here are two paragraphs from Chapter 7, "Programs and People."

"The asynchronous executive had one major drawback: the program flow was unpredictable.  No one could say exactly how it would respond in a given circumstance, because not everything was scheduled in advance; subroutines ran in real-time in response to events or as they rose up in the priority list.  One might make the analogy between the asynchronous executive in the Apollo computer and the Apollo "program" overall: numerous different tasks, working cooperatively but competing for attention and resources.

"Essentially a series of equations and flowcharts, the GSOP [guidance systems operations plan] theoretically contained everything required in advance to produce the program code for each mission.  In practice, the GSOPs tended to document code after it was written, rather than specifying it beforehand.

It all sounds so familiar and up to date.  After reading this book I wondered how much we've advanced in software projects and operating-system technology.  We have made strides, but I bet we could read critiques of many current projects that would sound much like the second quoted paragraph above.

I recommend this book if you have an interest in the development of some of the early embedded systems--ones that had no dev kits or integrated development environments.  Programmers wrote statements on punched cards and used a mainframe to simulate the processors still under development.  These processors had no flash memory or semiconductor ROM. Instead, wires woven by hand through magnetic cores creates a true hardwired "rope" of code.

I have enjoyed this book and think anyone with an interest in historic computer developments will enjoy it, too.  Younger engineers who were born after men left footprints on the Moon will better understand how we got to today's technologies--and how some of the same problems--management and technical--from almost 50 years ago still vex us today.

--Jon Titus

Book Note:
Mindell David A., "Digital Apollo: Human and Machine in Spaceflight," MIT Press, Cambridge, MA. 2008. 359 pages.
ISBN: 978-0-262-13497-2 (hardcover).