ming. AT&T Long Lines built the long-distance
communications; AT&T Western Electric built the
control centers; and Burroughs Corporation solved
analog/digital conversion by building modems.
The cost of Project SAGE was enormous. In addi-
tion to the expense of R&D and the deployment of
infrastructure, each Sector Direction Center oper-
ated four shifts per day of more than 100 people
per shift. Additional thousands were employed in
the command centers, the radar stations, the inter-
ceptor aircraft wings, and the missile batteries.
Was Project SAGE worth what it cost from a
defense perspective? It is difficult to know. The
ideal outcome was not air defense but deterrence:
making the risk of defeat in an attack so high
that the enemy would never attack. It is impos-
sible to know what role Project SAGE played in
the Cold War, but there never was an attack from
the north. Project SAGE ran for two decades and
never saw a bogie, much less a bandit. Moreover,
the intercontinental ballistic missile (ICBM)
introduced in the late 1950s made traditional air
defense moot: ICBMs flew so fast there was little
warning time before they struck, and they were
almost impossible to shoot down. Project SAGE
was kept in operation to protect against attack
by manned bombers. It was surely targeted by
Soviet ICBMs, but nuclear war was avoided and
Project SAGE never faced the ultimate test.

Leaving aside the defense outcomes of Project SAGE, it is fair to say the enormous investment in the effort paid off in the creation of a unique U.S. computer and communications industry that has generated hundreds of billions of dollars in value over the past half-century. Two SAGE contributions were directly related to HCI: interactive CRT displays and light-pen I/O devices that foreshadowed the creation of the mouse. In addition, SAGE introduced a number of important technical innovations: ferrite core memory used in computers from 1953 until the mid-1970s; analog/digital and digi-tal/analog conversion through the modem; multi-processing; real-time database management; distributed processing; time-sharing; marginal checking for component failure (arguably the beginning of intelligent devices); memory cycle-stealing

(LISP machines subsequently made this popular); buffered I/O; the COMPOOL (shared memory for subroutines that reappeared in COBOL); and large-scale system executives (the SAGE real-time executive had more than 500,000 lines of code). The

• A late 1950s SAGE console with “light gun” used to select radar tracks for display on central “summary board” (not shown). The operator access-es information and directs action using sliders, switches, buttons, and dials. Courtesy IBM Archives.

 

interactions

the north, integrating that information with other incoming information and known conditions, calculating the appropriate air-defense response, and launching counter-measures. Nothing like this had ever been tried before. Cooperation between the Canadian and American governments had to be established within NATO parameters, and a mix of military, public, and private organizations had to be brought together to work as a collaborative whole. MIT’s Lincoln Laboratories spun the Whirlwind computer off to the IBM Corporation as the IBM A/N FSQ7. The computer had 32-bit word length, four index registers, a real-time clock, magnetic (ferrite) core memory, and 60,000 vacuum tubes. It weighed half a million pounds and consumed three megawatts of power. Two A/N FSQ7s were built for each center, one of which would be in operation and the other of which would be in hot standby mode at any given time. MIT also spun off a new nonprofit organization called the MITRE Corporation to do systems integration. Other partners included the strategic-defense think tank RAND Corporation, which, with IBM, set up the Systems Development Corporation to do program-