Design considerations are presented for the construction of an operating system. (This approach emphasises the need to understand the interrelationships among all operating system components to facilitate the inevitable growth and development of a useful system). The concept of a process and a resource are defined in such a way that a resource-to-process concept is developed, which is later used to provide an ‘individualised’ virtual environment for a process. Terminology is introduced to distinguish between so-called S and type M resources, which must be treated differently for resource sharing. The levels of abstraction are described for the tree structured operating system that evolved from the design considerations. Within the tree, resource sharing is allowed and controlled by a communication language and so-called resource tables. Since the system was designed for use on a PDP-11/45, the basic components of its tree are given. It is then suggested that the tree presented here is, in fact, a basic framework upon which different specific operating systems may be built.

Operating systems have come into wide use within the last decade; many of today's systems have evolved into large, hard-to-understand complexes from small, workable monitoring systems. The object of this paper is to present the basic concepts of a well-structured environment which can support a highly complex operating system, but which at the same time can remain generally comprehensible, by virtue of its structure and the basic components of that structure. The paper begins with the basic design criteria and goals. It then presents the idea of a well-defined tree structure using the concepts of local-level knowledge (which delimits the scope of the processes at each node of the tree), modularity (which separates the system into workable parts), and functionality (which demands that each level of abstraction be well-specified). Two concepts, namely the process selector tree and recordability tree, are discussed as natural extensions of a tree structured system.

This paper presents a new model for use in scheduling processes for the sharing of a processor. The model may be used in various modes of operation, including multiprogramming, real time and time sharing. Because the model is applicable for various modes of operation and because it is significantly useful only within a well-defined system hierarchy, the context for discussion is Hansen's system for the RC 4000.3,4,5