> IIRC the noise you refer to as "key click" was actually the noise of > the solenoids banging the punches through the card. They had to be > strong and fast to keep up with an operator that could punch 10,000 > strokes an hour or better.
We had around 140 keypunch operators at any time working an early and late shift. The data was mostly numeric - credit card purchase transactions and remittances. The remittance was accompanied by a 51-column torn off statement that could sometimes be gang-punched - inbound remittances were split into 'minimum payments', 'whole balance payments' and 'odds', with the first two being gang punched from amounts prepunched in the returned segment. The odds had to be keypunched. It was pure numeric data, punched and then verified. I remember having a serious problem because we calculated keystroke rates for the operators - it happened that girl with a passing resemblance to Helen of Troy was going out with one of the computer room operators, who bore a reasonable resemblance to Adonis, and won the league every month. There were accusations of fiddling from the other girls, and I had to check the figures several times - she was watched for some time as well. She hit 26,000+ key depressions an hour with a card error rate less than 3%. Such measurements are now illegal in Europe. One problem I remember was that IBM stopped using two-shot mouldings for keytops and the operators would wear off the inscriptions in a week. It worried me, but of course _they_ knew which keys were which. We had to lean pretty hard on IBM to get them to find us keytops that were two-shot moulded - i.e., the figures were cast right through the keytop in two colours of plastic so it didn't matter how much they wore - you could still read them. -- Phil Payne The Devil's IT Dictionary - last updated 01/12/05: http://www.isham-research.com/dd.html
