1987
The Director's Keyboard
by Bob Graham
With a field such as ours, with a half-life of five years - where half of what we know today will be obsolete five years from now, and half of what we will need then hasn't even been invented yet - there is a large body of knowledge that has been replaced over the years. Rather than just let all this go the way of the passenger pigeon, some of the early things need to be written down before they are totally forgotten.
By default, this author has become the historian for the department, so I'll use this column to not only cover what looks like is coming in the future, but to also write about some of the more interesting things of the past.
The IBM 704, which is the machine that really made IBM, was a 36 bit machine. The object deck was punched in binary, three columns to make a 36 bit word. Compile times for a large program could run an hour or more, so patching a program was a real necessity. To patch a program you first figured out the machine language instructions needed for the patch and converted them to binary. The next step was to multipunch the binary into the three card columns for each word. This worked fine where you had to "add" pukas to the card, but when you had to change a one to a zero was another matter. You went to the keypunch and got some "chad", inserted one into the offending puka, and then rubbed it with the blunt metal end of something like a red pencil. This caused the chad to expand slightly and would stick in the hole firmly enough to go through the card reader.
The worst thing that could ever happen to you was to take such a patched deck, riffle the cards and have a "bit" go flying. It was enough to make any programmer cry.