Rick Beach, February 15, 1987 7:51:33 pm PST
ACM SIGGRAPH Workshop
on Software Tools for
User Interface Management
At a 1982 SIGGRAPH-sponsored workshop on Graphical Input and Interaction Techniques, the basic concepts of User Interface Management Systems (UIMS) were articulated. At a 1984 workshop in Seehiem, Germany, several such systems were reported and some requirements and issues further elucidated. In November of 1986, the ACM SIGGRAPH Workshop on Software Tools for User Interface Management was held at the Battelle Conference Center, in Seattle, Wash. The purpose was to synthesize new ideas and direction for future research through discussion and interchange. As in the previous two conferences, only 30 to 40 participants were invited. Participants were chosen to include a breadth of expertise and perspective, from experienced researchers to new graduates or current graduate students.
Care was taken not to weight the participation towards UIMSs because it was not clear that the UIMS concept or structure was still valid after four years. Many questions existed. Can a UIMS approach handle the needs of direct manipulation style interfaces? Where does a UIMS fit relative to operating systems which use graphical windows as a model for controlling multiple processes? Why are there not more commercial UIMSs? What is the relationship between a UIMS and a ``user interface toolbox''? Have our user interface tools solved problems that are different than the ones that the real world has? How do we expand the capabilities and power of user interface software?
Prior to the workshop, all attendees were required to submit position papers suggesting directions and issues in developing software for graphical user interfaces. These position papers were distributed prior to the conference, allowing participants to prepare for discussion. After a day of presenting information and selecting important issues, the workshop divided into four working groups. Each of these groups has prepared a report synthesizing their ideas. The reports are printed here, along with several original position papers.
The first group was concerned with the goals and objectives of user interface tools. In doing this the entire process of user interface development was explored using the software development lifecycle model. There were, however, a number of needs and issues identified which are unique to the problems of developing graphical user interfaces.
The second group concerned itself more closely with the process of designing user interfaces and on the tools that interface designers need. In many ways their issues echoed those of the goals and objectives group but from a different perspective. In particular this group spent more time on the integration of human factors issues into a designer's tools.
A third group addressed the software environment in which user interface tools must reside. Earlier models for user interfaces have considered monolithic applications which have total control over the graphics workstation. As workstations have developed whose graphics are integrated into multitasking facilities of the operating system this model has suffered severely. This group's report contains a model for how the hardware devices, operating system, windowing system, dialogue management and interactive applications all interrelate. This report also attempts to account for distributed interfaces, collaboration among users, concurrency and multiple kinds of interactive media which have not been clearly addressed in earlier work.
The fourth group took on the question of what the internal model of UIMS-supported applications should be. Of particular concern was the relationship between the UIMS's dialogue management functions and the application. As more direct manipulation interfaces are being designed the concept of rigidly separating the application from dialogue management starts to break down. This is a particular problem as semantic feedback becomes a more dominant part of the interactive process. On the other hand to totally integrate the application code with the dialogue management is to forfeit some the potential of a UIMS. These tradeoffs and others are the subject of this group's report.
One issue which was not addressed in any of the working group reports is the issue of standardization. There was simply too wide a range of opinions for any synthesized position. There are the obvious desires for a standard around which to unify software development. There were several expressions of standardization being the formalization of anachronisms. This pessimism stems from the fact that changes in the graphics world have been so rapid as to render many standards partially obsolete before they are completed. Others expressed the fact that the user interface field is not mature enough to build a really good set of tools let alone produce a standard set.
The reports presented in this issue form a basis for future research and a strawman to beat upon for the next few years. More importantly the workshop provided an opportunity, not readily available in a conference the size of SIGGRAPH, for researchers to come together and exchange ideas and to expose their opinions to active, rather than passive or indirect, analysis and debate. The three workshops have been a major catalyst in user interface management research. When all is said and done it is people not machines, software or organizations that make the contributions in any field. This workshop was a significant investment in those people.
Goals and objectives group:
Bill Betts, Inference Corporation
David Burlingame, Software Technology Center, Lockheed Missiles
Gerhard Fischer, Dept. of Computer Science, University of Colorado
Jim Foley, Dept. of Electrical Engineering & Computer Science, George Washington University
Mark Green, Dept. of Computing Science, University of Alberta
David Kasik, Boeing Computer Services
Stephen T. Kerr, University of Washington
Dan Olsen, Computer Science Dept., Brigham Young University
James Thomas, Battelle Pacific Northwest Laboratories
User interface designers group:
John L. Bennett, IBM Almaden Research Center
Teresa W. Bleser, Dept. of Electrical Engineering &Computer Science, George Washington University
Neil B. Corrigan, Westinghouse Hanford
Roger Ehrich, Computer Science Dept., Virginia Tech
Kurt J. Schmucker, Productivity Products International
Thomas T. Hewett, Dept. of Psychology & Sociology, Drexel University
Przemyslaw Prusinkiewicz, Dept. of Computer Science, University of Regina
Henry Ramsey, Boeing Computer Services
James Rhyne, IBM Watson Research Center
John Sibert, Dept. of Electrical Engineering & Computer Science, George Washington University
John Wiberg, Battelle Pacific Northwest Laboratories
Reference models, windowing and concurrency group:
Carl Binding, Dept. of Computer Science, University of Washington
Andrew Dwelly, ECRC, West Germany
Kuan-Tsae Huang, IBM Thomas J. Watson Research Center
Keith Lantz, Dept. of Computer Science, Stanford University
Peter Tanner, Dept. of Computer Science, University of Waterloo
Runtime architecture group:
John Dance, Battelle, Pacific Northwest Laboratories
Tamar Granor , Dept. of Computer and Information Science, University of Pennsylvania
Ralph D. Hill, Computer Systems Research Institute, University of Toronto
Scott Hudson, Dept. of Computer Science, University of Arizona
Jon Meads, Jon Meads Associates
Brad A. Myers, Dept. of Computer Science, University of Toronto
Andrew Schulert, Apollo Computer Inc.
Henry Lieberman, Artificial Intelligence Lab., Massachusetts Institute of Technology