Showing papers on "User modeling published in 1980"

Uniform interfaces for distributed systems

Abstract: At the University of Rochester we have had five years of experience in the design and implementation of a multiple machine, multiple network system called RIG. It has been useful to distinguish two types of system service: The first is provided directly to the user through a user interface, or user-system interface. The user obtains these services by typing commands or requests which are satisfied by actions initiated by the user interface. The user interface should provide a common viewpoint for the user of distributed resources. The second type of system service is provided by programs or processes executing on behalf of the user. The user's program obtains these services by executing "system calls." The system software that interprets and satisfies these calls implements the system interface or program execution environment. In RIG, equal emphasis was given to user and system interfaces. Each was made as uniform and coherent as possible across all available resources. It is the purpose of this thesis to present certain mechanisms and paradigms for building distributed systems with uniform interfaces. Building on a base of processes communicating solely via messages, four key contributions to research in distributed systems are described: (1) Virtual Terminals are presented as the means for managing a large number of application programs per user. The Virtual Terminal Management System provides extensive facilities for editing text, the ability to save all output in disk-based data structures, and sophisticated mechanisms for the management of screen space. (2) Principles of command interaction are outlined which facilitate the use of distributed resources. Together with the Virtual Terminal Management System, the command interface serves to present an elegant, robust, and consistent interface between RIG and the user. (3) Mechanisms for resource management are presented which are well suited to a distributed environment. In particular, resource management is viewed fundamentally as a problem of process management. Processes may be created "by name" and registration facilities enable any process to register its interest in, for example, the death of any other process. (4) Paradigms are presented for how processes should be written and communicate. Of particular importance are the simple yet powerful mechanisms for error and exceptional conditional handling. Registration facilities are provided which enable any process to register its interest in exceptional events which occur with regard to any other process.

Semantics vs. graphics - to show or not to show

Abstract: This paper describes a high-level graphical interface to databases which employs specialized knowledge about graphical symbols, the underlying database, and the database user's context in order to provide pictorial representations of data which is of concern to the user. The approach combines techniques of graphics and knowledge representation to provide more complete and more perspicuously organized answers to user queries. Techniques are presented for the representation and use of the semantic knowledge about the database contained in the data model. The knowledge of the data model allows the system to augment the response to a query with information which is relevant but not specifically requested. It also enables the data to be organized in a manner which reflects the relative importance of each of the entities and attributes within the context of the user's question. Thus the answer to the user's question is broadened in a meaningful way to provide the user with a more complete perspective of the information in the database.

User needs, the new technology and traditional approaches to library services

Abstract: This paper indicates weaknesses in the concept of user needs as a concept of practical utility in the task of identifying the right book at the right time; it doubts the value of an 'active' approach to service in libraries; it comments on the development of online services in relation to 'user needs' and suggests a more sophisticated approach to record creation.

Perspectives of the user interface

Abstract: It has become increasingly popular for researchers to study management information system users and their relationships with system developers. Motivated by numerous reports of system failure, which were not technical in nature, investigators have sought a better understanding of the user interface. The ultimate objective of much of this research is to guide methodologies for user-oriented system development and design of systems which meet user needs more readily. In other words, the intent is to design systems which are behaviorally valid as well as technically valid.Such a quest necessarily requires us to draw upon concepts from the behavioral sciences. To study user behavior, interpersonal relationships and organizational settings, one must either invent new theories or use existing ones. The latter strategy is strongly recommended for many reasons not reiterated here. But within the behavioral sciences there are fundamental choices among perspectives, and these choices influence our view of the user interface quite profoundly.Adopting any one approach allows the investigator to see and understand certain phenomena but to ignore others. Like the blind man examining the elephant, each observer develops a “trained incapacity” to see the object from other viewpoints. As a result, alternative explanations of research findings are rarely considered, much less reported.The purpose of this paper is to examine four different perspectives of the user interface: (1) user motivation, (2) user-developer differences, (3) organization structure, and (4) the political perspective. Each highlights certain legitimate system development issues. However, each also narrows our view of the user interface. Awareness of these perspectives accomplishes two purposes. First, researchers will be more receptive to the alternative perspectives and aware of the consequences of their choices. Second, investigators will treat the user interface less naively, especially through increased sensitivity to the political factors in system development.

Tools for constructing user interfaces

Abstract: A review of the tools available for easing the tasks of constructing and modifying the user interface of CAD application programs.

EP-2: An Exemplary Programming System

Abstract: : This report describes the design and implementation of EP-2, the current version of a prototype exemplary programming system developed as a programming aid for non-expert computer users. The EP system allows software to be created by example, i.e., the user performs some task on a computer while EP- 2 'watches,' recording the interaction. EP-2 then constructs a model of interaction and stores it for subsequent use. The system can recognize various prompts that applications programs send to the user, construct sequential actions under the user's direction, and use portions of program output in subsequent inputs. An agent library stores, indexes, describes, and retrieves agents the user has created, and editing facility enables errors to be corrected during agent creation. An experimental program-specification facility is also included that provides a 'scratchpad' for specifying functions to EP-2.

Infrastructure Approach to Integrated CAD Systems

Abstract: This paper is an introduction to the concepts and terminology now relevant in what may be called design methodology. I have not tried to say what is good or bad with specific designs but rather have tried to give general comments and rules of the thumb which experienced designers have found to be useful and successful. The framework which is proposed can be used to follow during the design process and to incorporate the aforementioned ideas and observations.