see "Notes on the Evolution of Acceptance Sampling Plans: Part I" By H.F. DODGE JQT, ASQ 1969:
In the late 1020's Dr Shewhart was working with Western Electric. At that time newly-designed, machine-switching
central offices were being installed, each with several thousands of lines. Each office was a highly complex
assembly of apparatus and wired equipment of various kinds, hundredreds or thousands of each of
several specific types of relays, switches and other kinds of apparatus. Before turning a central office
over to the telephone company, it was standard practice to (1) conduct a wide variety of routine
tests to prove in the functioning capabilities of the office, and (2) inspect the physical condition of apparatus
with respect to its many functional and adjustment characteristics. The apparatus had been
closely inspected and adjusted for conformance to specifications at the factory, but there was need for
further verification of its quality following its shipment to the central office site and its installation and
adjustment by installers.
Probability-of-Acceptance Curves
Working with the Installation Department engineers who had initiated this particular application,
the group took up the problem of how best to sample the different populations of relays and other types of
apparatus in an individual office. In these studies scattered sampling plans were selected and charts
drawn to show what were then called "probability of-acceptance curves", but later came to be known as
"operating characteristic curves" or "OC curves".
For an individual sampling plan, this curve showed the probability of acceptance plotted against lot
quality expressed in terms of per cent defective, and indicated how the plan would operate for different
values of lot quality that might be presented to the inspector.
------------------------------
Gregory Gruska
------------------------------