The traditional object paradigm is evolving into a paradigm of distributed, self initiating agents. To specify the functional requirements of agent systems, we need languages capable of analytic and coherent description of self initiation. The paper assesses the suitability of SF relative to other approaches, using criteria that consider the requirements of traditional software engineering and interactive systems development. The paper shows that SF has important advantages and suggests some directions for future work.>

This survey reviews four classes of reengineering of software systems: code restructuring and translation of code into a new language, design recovery for corrective maintenance, design recovery for software enhancement, and software reengineering as part of radical business process reengineering. Special attention is given to the reengineering of data bases and user interfaces. We also consider concurrent software engineering, its place in the software reengineering process, and groupware in support of this process.

A system is safety-critical if failure of the system would result in loss of human life, personal injury, or significant material loss. Software that supports or supplants human control of safety-critical systems has to be highly reliable, but much research remains to be done before all the reliability-related issues of safety-critical software are fully understood. We discuss a research agenda under the following headings: completeness of requirements, readability of specifications, validation tools, validation of responsiveness, verification of implementations, software robustness, common-cause failures, reuse of reliable software, the software process, and ethical issues.

We define the query language Vizla, in which answers to queries are built up by pointing to representations of sets and functions in a conceptual model of the data base of an application, and to iconic identifiers of computational operators or control constructs. At its present stage of development the primary use of Vizla is in the validation of conceptual models of information systems, but it is to be developed into a user interface to SF, a prototyping language for information and control systems. Moreover, it can be regarded as a visual programming language in its own right. As such it is based on abstract data types.>

Shows that data and knowledge base integrity constraints can be regarded as queries. This allows the author to use his visual query language Vizla for the construction of constraints in an operational style. The primary tool for this is the Vizla filter, which lets through only those elements of a set that satisfy a filter predicate. The main contribution of the paper is a method of using filter predicates as constraints, but fuzzy integrity constraints are also considered.>

Abstract Tagging of objects produced by generators is a simple mechanism for the synchronization of generators. Such tagging is proposed as a reliable software design technique for problems based on data streams. It is used to develop programs for two‐way merge and file update.

Theoretical results have had much greater impact on computing practice than we are normally aware of, and the practical significance of theoretical results can be expected to become more prominent in the future. We discuss the past and present significance for data processing specialists of some results in analysis of algorithms, languages, and program proofs, and argue for a greater emphasis on computer science theory in data processing curricula.

The need for graduates from master's programs in computer science and related areas is well recognized [19, 20]. Indeed, some companies have a policy of extensively utilizing master's programs at universities for the continuing education of their employees. The Graduate Study Program of Bell Laboratories is well known. At Honeywell Information Systems it has been found that support of continuing studies at the master's level helps in hiring and retaining personnel, and is beneficial to the dissemination of new technology through the organization [29]. It has been demonstrated that programmers acquire new knowledge primarily from other programmers [17]; periodic influx into an organization of graduates of programs of advanced study is therefore essential if the organization is to retain technical soundness.Universities have responded to this need, but in a rather haphazard manner, with the result that we have today a variety of programs, some of which have very little to do with computer science. Some of the programs are no more than, to use Smoliar's [41] words, “undergraduate programs for grown-ups.” Others are viewed as a first stage in the preparation for research careers of narrow specialization. Late in 1972 Terry Walker [45] conducted a poll of master's degree granting departments. The four primary objectives of a master's program given by the 93 respondents were: prepare a person for a job designing computer software systems, prepare a person for a job as a systems analyst, prepare a person to pursue a doctoral degree in computer science, prepare a person for a job as a scientific programmer. Today one would add a fifth objective: prepare a person for teaching computer science at the junior college level. There is clearly a need to reconcile these different objectives with a unified view of computer science.

Eight years after Curriculum 68 there is still much uncertainty regarding the place of discrete mathematics in a computer science curriculum. This paper explores the purposes of discrete mathematics courses in a very broad context. It is suggested that discrete mathematics be distributed over two courses, and that the attitudes established in the courses are more important than the contents of the courses.

A reasonably efficient procedure for testing pairs of directed graphs for isomorphism is important in information retrieval and other application fields in which structured data have to be matched. One such procedure, a backtrack procedure based on a representation of directed graphs by linear formulas, is described. A procedure for finding a partial subdigraph of a digraph that is isomorphic to a given digraph is also described.