The Apollo Guidance Computer

Mind-boggling material on the Apollo guidance computer hardware and software, from David A. Mindell's Digital Apollo: Human and Machine in Spaceflight.

"Battin broke the problem down into 'decision points; on a lunar flight: an observation is made, the state vector is updated, and then either a correction is made (e.e. a rocket is fired) or no action is taken. A set of decision rules control the number and frequency of these observations. Each measurement would update the state vector in some way, depending on the statistical level of uncertainty... in the measurement...[Battin used] a catalog of stars and planets; at each decision point it would suggest which combinations would provide the best new information. In a typical Earth-moon trajectory, taking about 62 hours, Battin developed a plan for 41 observations, resulting in four velocity corrections...the position uncertainty when arriving at the moon would be 1.2 miles...enough for a precise orbit."