This paper follows a trend towards more user oriented design approaches to interactive computer systems. The implicit goal in this trend is the development of more “natural” systems. Design approaches should aim at a system capable of continuous change by means of suitable agents. The term “soft facade” is used to describe a modifiable interface to a system. Facades vary in softness and agents for change can be the systems designer, the user or an expert system in the facade. In the latter case, the system is called a self-adaptive interface system. The conditions where a self-adaptive interface system is desirable are briefly considered and discussed in relation to a simple example. Recent research in artificial intelligence is mentioned as a possible source of proposed components for a self-adaptive interface system.

Towards self-adaptive interface systems

Research issues concerning human interaction with computers have become crucial because of dramatic increases in the number of computer users who have no formal mathematical, engineering, or computer science training. This review focuses on the state of the art in Software Psychology--the study of human factors in computer systems. One area of software psychology, the psychology of programming, is picked for an in-depth discussion of research results. This area was selected for its interest to the readers of the PROCEEDINGS OF THE IEEE. The review then discusses the methodological issues involved in performing experimental research in software psychology. Future trends are discussed both for the programming research reviewed here and for the whole field of software psychology. The conclusions identify multidisciplinary training and research as critical to further progress in this field.

Software psychology: The need for an interdisciplinary program

Knowledge-based dialogue for dynamic systems

A message based model of a Distributed Decision Making, DDM, System is presented. This model is developed from a survey of the needs of group decision-making processes within organisations and distributed computer applications; a review of this survey is given. DDM extends the DSS concept to one that enables two or more decision making parties to co-operate in using DSS tools. It assumes a framework of organisational decision making which concentrates on the organisation as a political system where communication is important and conflict between groups just as likely as consensus. A brief description of a prototype implementation of the model is given. The use of the prototype has upheld the principles from which DDM was designed.

A distributed decision-making system

The CONNECT interactive dialogue system is briefly described and the path-algebra technique explained. Its relevance to interactive dialogue design is illustrated by a number of examples. The analysis and usefulness of multilayered networks is briefly explained and a possible extension to step retracing is outlined.

The application of path algebras to interactive dialogue design

The development of computer programs, called agents, that act as man-machine interfaces for computer users are described. These programs are written in RITA: the Rule-directed Interactive Transaction Agent system, and are organized as sets of IF-THEN rules or “production systems.” The programs, or “personal computer agents,” are divided into two main categories: those that interface the user to computer systems he wishes to use and those that interact with the user to acquire the knowledge needed to create these interface programs. The relationship between the interface program and the knowledge acquisition program is that of parent-offspring. Three types of parent-offspring RITA agent pairs are described: (1) an exemplary programming agent that watches a user perform an arbitrary series of operations on the computer and then writes a program (a task agent) to perform the same task; (2) a tutoring agent that watches an expert demonstrate the use of an interactive computer language or local operating system and then creates a teaching agent that can, help naive users become familiar with the language or system demonstrated by the expert; and (3) a reactivemessage creating agent which elicits text from a user (the sender) and from it creates a new RITA agent which is a reactive message. The reactive message is sent to some other user (the recipient) who interacts with it. During the course of the interaction a record of the recipient's responses is sent back to the sender.

/pdf/a-rule-based-approach-to-knowledge-acquisition-for-man-1kyde9sx05.pdf

D.A. Waterman

Papers

A rule-based approach to knowledge acquisition for man-machine interface programs