Apollo Guidance Computer History Project

Second conference

September 14, 2001

Herb Thaler's introduction

HERB THALER: I got involved in the Apollo project by way of sort of a back door. I was working for Raytheon and attached to the Instrumentation Lab on the Polaris project as a resident. Shortly after that petered out a little bit, I became reattached as a resident on the Apollo program in the computer group. Just putting things in context, I was just all of three years out of school. So not what you'd call really tremendously experienced in computer design. But got into the computer group and was doing work with Ramon and Dr. Hopkins, Dave Shansky to go ahead with the Mars computer.

At then shortly after arriving, the concept of the integrated circuit arose. I remember meeting in Eldon Hall's office where we all gathered around and tried to determine which direction the group ought to go as far as implementing either with the Mars circuit technology, the core transistor logic or the integrated circuit approach and all of the issues that that was going to raise. And the decision was made to go ahead with the integrated circuit approach. Of course none of us knew what an integrated circuit was at that time.

I ended up with the responsibility of reimplementing the Mars design in the integrated circuit world and carried that through. It's a very interesting computer design. Many levels of interrupt. I remember words like pink, mink, shink and shank for essentially instantaneous reaction to requests from resolvers and synchros and communication links and things like that, and the multi-structured level of interrupts and software levels of hierarchy.

Hugh Blair-Smith adds:

The ones mentioned are

PINC = Plus INCrement: in one memory cycle, add 1 to the contents;

MINC = Minus INCrement: in one memory cycle, subtract 1 from the contents;

SHINC = SHift INCrement: in one memory cycle, shift the contents left by 1 bit;

SHANC = SHift & Add INCrement: like SHINC but inject a ONE into the vacated bit.

SHINC and SHANC will be recognized as the receiving function of a UART (for uplink). The output (downlink) function was achieved, naturally, by OINC. There isn’t any actual incrementing going on in those three, but we applied the designation INC to any of the DMA cycles that took advantage of the two-step operation of core memory: readout of the prior contents about one third through the cycle, and writeback of the new contents about two thirds through.

I also remember a fateful day in 1963 sitting in the Apollo computer lab which was a big room and then a little room behind it. I was in the little room behind it with the first brass board Apollo computer prototype. I was debugging it when the word came that Kennedy was shot. That was kind of a real seminal moment. I think everybody became particularly re-energized to make this project work. That was my role.

I did a little bit of programming when the design was finished to some orbital calculations. But my role was primarily in the design of the hardware.

Organization of the AGC project

site last updated 12-08-2002 by Alexander Brown