Contemporary Cybernetics

COMPARISON OF COMPUTERS One of the real problems of the day for the research worker, who wishes to acquire a new computing facility, is the problem of the evaluation and testing of computing systems. Obviously the best method of comparison of software or hardware is via the Formal Description. Languages of formal description have been developed for programming and also for computer architectual comparisons. Such a notation, ISP, was developed to formalize the information normally given in basic machine manuals and to supplement and, possibly replace programming reference manuals. This notation was introduced for descriptive purposes by Bell and Newell (Computer Structures, McGraw Hill, New York) in 1971. For actual use it was revised and renamed ISPL, with further developments during the mid and late seventies. In a project at the Carnegie‐Mellon University, Drs. Barbacci and Siewiorek used Formal Computer Descriptions for the evaluation of test programs. The formal language ISPL was used to define the architecture of a computer. The description compiles into an intermediate form that can be linked to and interpreted by a simulator. A symbolic translator converts assembler‐generated outputs of test programs into commands that initialize the simulator. After execution of the test program a post‐processor summarizes the dynamic behaviour. This is surely the first time that the architectures of commercially viable computers have been described in formal language, descriptions compiled and then used to drive a simulator executing benchmarks, and diagnostic machine language programs.