"In 1948, at the request of the computer group at Northrop Aviation, Inc., IBM coupled a 603 calculator with a 405 accounting machine, using the 405 as a control unit and printer. This combination was so successful that after the 604 came out, Northrop approached IBM with a proposal and request for a more powerful version based on the 604. IBM responded by developing the card programmed calculator (CPC).
The CPC was made up
of three units: a 605, a 417, and a 941. The first two were modifications
of the 604 calculator and the 402 accounting machine, respectively, and
the 941 was a special box containing 480 digits of additional storage using
the same type of electromechanical storage used in the 417. Programs of
almost any length and complexity, punched into IBM cards, could be fed
through the reader unit of the 417, with the mathematics performed by the
605, the intermediate results stored in the 605 or 941, and final results
punched into cards in the 605's punch unit, or printed by the 417 print
unit. Almost 700 of these combinations were built. As Rosen wrote (1969),
The CPC was slow by electronic computer
standards, running at a maximum speed of 150 instruction cards per minute.
Yet it is hard to exaggerate its role as an interim computer, carrying
the major computing load in dozens of computation centers while they were
waiting for the stored program computers to live up to their promises in
terms of delivery and performance." [Annals of the History of Computing,
Vol.2 #3 July 1980]
Personal Reflections - Roger Mills: