The concept of data abstraction is utilized in database systems to define user interfaces via database views in database application languages and to describe the architecture of database systems. Differences between the specification and use of database views and other data abstractions realized as abstract data types are discussed. Database views are formally specified using both the algebraic specification method and the abstract model specification method. The use of database views is demonstrated via the EXT_Pascal database application language.

The use of database views in database management systems that enforce user level discretionary and nondiscretionary access control policies is discussed. This discussion involves several issues such as how should views be classified?, what types of mechanisms should be used to define views?, etc. Mapping between views, view updating, and aggregation and inference problems are also discussed.

A specifilcation method designed primarily for specifying data abstractions, but suitable for specifying procedural abstractions as well, is described. The specification method is based on the abstract model approach to specifying abstractions. Several data abstractions and procedural abstractions are specified and a proof of implementation correctness is given for one of the data abstractions–a symbol table.

A data structure definition facility (DSDF) is described that provides definitions for several primitive data types, homogeneous and heterogeneous arrays, cells, stacks, queues, trees, and general lists. Each nonprimitive data structure consists of two separate entities—a head and a body. The head contains the entry point(s) to the body of the structure; by treating the head like a cell, the DSDF operations are capable of creating and manipulating very general data structures. A template structure is described that permits data structures to share templates. The primary objectives of the DSDF are: (1) to develop a definition facility that permits the programmer to explicitly define and manipulate generalized data structures in a consistent manner, (2) to detect mistakes and prevent the programmer from creating (either inadvertently or intentionally) undesirable (or illegal) data structures, (3) to provide a syntactic construction mechanism that separates the implementation of a data structure from its use in the program in which it is defined, and (4) to facilitate the development of reliable software.

The course and project description for a one quarter (or semester) senior level course in evaluation and measurement techniques for computer systems is described.The primary purpose of the course is to introduce the students to evaluation and measurement techniques and to use some of these techniques in evaluating and measuring the performance of a multlprogranuning operating system designed for a simple machine'~ The proJect is based on the availability of a simple multlprogrammlng operating system that can be modified by stydents to perform evaluation and measurement experiments.Five experiments are described in the pat per.If a simple multiprogra~uing operating system is not available then the project may be omitted.

Analytical results in terms of Laplace-Stieltjes transforms for the busy period distribution of a time-sharing system with a central processing unit (CPU) and N peripheral units are derived in an earlier paper by Bhat and Nance.Assumptions of the model are that job arrivals are Poisson, required processing time for each job is an independent exponential random variable, and a feedback type discipline governs CPU access.The steady-state results of the earlier work are extended to gain information about the transient behavior of the CPU, Approximate results for the busy period distribution function are presented through numerical inversion of the transforms.Using these results qualitative conclusions are drawn regarding the sensitivity of system performance to changes in the number of peripheral units, quantum size, and CPU capability.