This paper (or summary?) appears on page 716 of:
Austin Curwood Hoggatt, editor. Proceedings of the 1973 Winter Simulation Conference. Sponsored by ACM/AIEE/SHARE/SCi/TIMS, January 1973. https://files.eric.ed.gov/fulltext/ED077182.pdf
The needs for laboratory control of man-machine interaction form a super set of the needs which lead to time sharing systems. In addition to facilities for on line construction and debugging of programs and operating systems aids to the allocation of resources, there are real time constraints on system response in order that subjects are placed in environments which facilitate the recording of observations on their behavior with minimal delays or machine induced artificiality of situations. Requirements for console interaction and control of physical facilities (tape recorders, t.v. monitors, special input output devices) require the development of an operating system which accomodates these needs. A need to describe models of the subject environment in an operational publication language led to the adoption of APL as the experimenter's language to which have been added extensions to allow parallel operation of instances of the same process, multiple I/O units under control of a single process, and control to provide for protection of the integrity of ongoing experiments. Reasons of economic viability dictated that this be accomplished on a low budget system. Under N.S.F. funding the CRMS Laboratory is pursuing the development of a pair of micro-programmed systems on which respectively are implemented the operating system and the APL engine. We emphasize the interplay between needs of experimenters, software and hardware design and the allocation of functions between hardware and software in order to minimize the difficulties of realization in the resulting system.