The design motivation for a real-time control nucleus is normally very process/goal specific. Based on a goal to minimize the number of primitives, this paper recommends a set of functions to be included in such a nucleus.

The paper provides a reasonable survey of process and data access control mechanisms used in general concurrent software systems. An interface cluster approach is proposed as the most appropriate for real-time control systems. This approach supports ACQUIRE and RELEASE procedures for access control and WAIT and SIGNAL for synchronization. The support for the recommended approach is, at best, qualitative.

The paper is well written. It is disappointing, however, that the authors do not either analytically or empirically justify their sweeping conclusions.