1970: Personal computing with a mainframe?
My first "personal" computer was an IBM 1401. Wait, I'm serious!
In 1970, I used a 1401 Data Processing System at a U.S. Navy supply depot in Vietnam during the war. The 1401 and its associated card reader, card punch, chain printer and tape drives filled a small room.
Although the system was obviously not intended to be used by just one person, we in effect treated it that way much of the time. It was like a time-share condo -- you had it for maybe just 30 minutes, but when you had it, it was yours.
Here's how it worked: The computer had just 12,000 characters (bytes) of magnetic core memory. Instructions and data were stored in memory in variable-length words separated by special characters called wordmarks and, unless you changed your program, each word went into the same, known memory location every time the program ran.
When your program was running, nothing else was running, not even an operating system, because there was none.
These characteristics, plus some handy-dandy user-controllable hardware features, provided the ultimate in debugging capability. You could stand at the console and, by manipulating a series of toggle "sense switches," step through your program a few instructions at a time.
At each key point as you stepped through your code, you could check the contents of registers via little lights to see if something got clobbered, and you could check the address of the next instruction to make sure that an unanticipated branch had not occurred. If you saw that a certain instruction wasn't working the way you wanted, you could just override it by inserting a "no-op" (a placeholder instruction that does nothing) and keep on running.
This could be a slow and tedious process, to be sure, but the cool thing was that you had complete knowledge and control over the execution of your program. Try that the next time you get a blue screen of death.
-- Gary Anthes