The possibility of widespread industrial application of this as well as many other control systems materialized with the electronics revolution of the 1950s and 1960s, which furnished cheap and reliable circuitry for the control instruments. This began with the transistor, which at first simply replaced the vacuum tube on a one-for-one basis. By the early 1960s, however, integrated circuits combined transistors and other components on tiny chips of silicon crystal, so that eventually large-scale integrated circuits offering the functions of hundreds of expensive and bulkier parts were combined on a single chip. As the yield of batch production processes climbed, the cost fell from an average price per circuit function (one transistor) of \$2 in 1965 to under three cents in 1971. The reliability of operation and ease of repair through modular replacement, combined with this cheapening of increasingly complex circuitry, are the basis for the revolution in control technology, and it is here rather than in earlier experiments that the source of this new industrial technique and its broad use must be sought.\*
By 1968, no more than 1 percent of machine tools in industrial use were numerically controlled, but the shape of the future could be seen in the fact that 20 percent of all new machine tools shipped in that year were equipped with this attachment. And in trade showings of machinery, as well as in industry journals, the big majority of machines displayed or advertised are now tools of this type.
With numerical control, the machine process is subjected to the control of a separate unit, which receives instructions from two sources: in numerical form from an external source, and in the form of signals from monitoring devices which check the ongoing process at the point of contact between tool and work. Using this information, the control unit originates signals which activate power drives controlling the work, tool, coolant, etc.
From a technical point of view, the system offers several advantages. Complex metal cutting—for example, the machining of surfaces to compound curves—slow and demanding when the calculations are made in the course of cutting, may be coded with relative ease and cut with assurance; this was one of the features which, because of its applicability to the shaping of dies and other parts used in aircraft production, interested the air force in the method. The coding of any job is quickly completed when separated from machine execution, and once coded a job need never be analyzed again: the tape may be kept on file and used whenever a remake is called for. The processes of metal cutting are virtually automatic, relieving the worker of the need for close control of the machine while cutting is in progress. The separation of conceptualization and calculation from the machine means that the tool itself is in more constant use for metal cutting; at the same time, it goes through its continuous cutting path without interruption, which also makes for more efficient use of these expensive pieces of equipment.
The unity of this process in the hands of the skilled machinist is perfectly feasible, and indeed has much to recommend it, since the knowledge of metal-cutting practices which is required for programming is already mastered by the machinist. Thus there is no question that from a practical standpoint there is nothing to prevent the machining process under numerical control from remaining the province of the total craftsman. That this almost never happens is due, of course, to the opportunities the process offers for the destruction of craft and the cheapening of the resulting pieces of labor into which it is broken Thus, as the process takes shape in the minds of engineers, the labor configuration to operate it takes shape simultaneously in the minds of its designers, and in part shapes the design itself. The equipment is made to be operated; operating costs involve, apart from the cost of the machine itself, the hourly cost of labor, and this is part of the calculation involved in machine design. The design which will enable the operation to be broken down among cheaper operators is the design which is sought by management and engineers who have so internalized this value that it appears to them to have the force of natural law or scientific necessity.\*
Labor and Monopoly Capital
Machinery
137
\* That engineers think in this fashion, or are guided in this direction by all the circumstances of their work, will not appear strange to anyone with the slightest familiarity with engineering as it has developed from its nineteenth-century beginnings. “The monogram of our national initials, which is the symbol for our monetary unit, the dollar, is almost as frequently conjoined to the figures of an engineer's calculations as are the symbols indicating feet, minutes, pounds, or gallons," said Henry R. Towne, the industrialist and pioneer of shop management, in a paper read to the American Society of Mechanical Engineers in 1886."The dollar," he said on a later occasion, "is the final term in almost every equation which arises in the practice of engineering. . . ."14 Or, in the words of a chemist in more recent years: “I'm no longer really interested in problems that don't involve economic considerations. I’ve come to see economics as another variable to be dealt with in studying a reactionthere's pressure, there's temperature, and there's the dollar."15
\* On this, see Business Week's recent special issue on productivity, which characterizes the goals of increased productivity through solid-state electronic controls as "fewer parts, less-skilled, low-cost labor, and fewer manufacturing steps."13