greenberg@calgary.UUCP (Saul Greenberg) (11/12/85)
Iconic interfaces for office systems based on video-games
Saul Greenberg and Roy Masrani
Man-Machine Systems Laboratory
University of Calgary, Canada
Abstract: Although users are enthusiastic about modern icon-based
interfaces using direct-manipulation principles (Shneiderman, 1983), we
argue that their full potential has not yet been realized. Specifically,
certain aspects of video-games are mergeable with conventional icon
systems (Malone, 1982), providing "players" with an exciting way of
getting their work done.
We have developed the WAP system (Work And Play) at our lab to test our
ideas. Icons are used as in standard systems, however icon manipulation
has an exciting video-game flavour providing an action-packed session.
For example, the passive trash-can is discarded and replaced with a
sophisticated lazer weapon icon. The "player" may aim the
weapon at a file and attempt to destroy it. Of course the file will take
evasive action, using other files as camouflage if necessary. To add an
element of risk, shooting the wrong file destroys it too.
After considerable experimentation, we found that a statistically
significant (p = .0001) number of subjects had poor aiming skills,
resulting in the loss of many years' work and corporate displeasure (Ed.
note: luckily we were able to salvage this abstract). To compensate, a
"God" Icon (introduced to the current WAP generation) is able to
re-incarnate dead files or to create new ones. In keeping with the
gaming spirit, a robust natural language interface uses "praying"
scripts (Schank & Abelson, 1977) to give users an opportunity to convince
God to restore their files. Proper user interface techniques, of course,
dictate that the God icon has a complete representation of the user and
may decide to shoot a lightning bolt at him/her (which re-initializes
the disc). We are now experimenting with a Devil icon as a counter-measure.
AI techniques will make mouse selections more challenging. Using complex
knowledge-bases, Icons will evade their selection. When "caught", icons
will be pierced by an arrow and squirm in agony. Of course, appropriate
sound effects are provided.
We have been experimenting with the system for the past 11 months. We
found that productivity has improved dramatically, as measured by login
hours. Unfortunately, funding is running low and we are searching for
sponsors.
References:
Malone, T. (1983) "Heuristics for designing enjoyable user interfaces: lessons
from computer games" Proceedings of human factors in computer systems
(SIGCHI), Gaithersburg, Maryland
Schank, R. & Abelson, R. (1977) "Scripts, plans, goals and understanding"
Lawrence Erlbaum Associates, Inc., Hillsdale, New Jersey
Shneiderman, B. (1983) "Direct manipulation: A step beyond programming
languages" IEEE Computer, 16(8)sanford@cdstar.UUCP (Conrad Sanford) (11/20/85)
> Abstract: Although users are enthusiastic about modern icon-based > interfaces using direct-manipulation principles (Shneiderman, 1983), we > argue that their full potential has not yet been realized. Specifically, > certain aspects of video-games are mergeable with conventional icon > systems (Malone, 1982), providing "players" with an exciting way of > getting their work done. While this interface model is an admirable motivational technology, I would like to point out one of the long-term considerations that seems relevant, if not imparitive. This being the type of emotional involvement implicitly engineering into the user interface. The video-game model raises a specific question that points to a more general one. Specifics first: Is it appropriate to increase a user's emotional involvement with a model that is filled with conflict? We should take it for granted that the psychological tool in making this form of interface a motivating one is the increased emotional involvement. What is striking is it's use of gods and devils, and the use of weapons and defenses. Previously purely intellectual tasks become tinted with some form of the potential for winning and very subtle personal identification with the characters of this model. The more general problem is thus: it seems likely that some eventual user interface will rely heavily on the user's emotional involvement, as well as his intellectual skills. I suspect that this will go beyond simple motivation, having an greater effect on his skill level. What factors should be implicit in *any* user interface model that would be of psychological and/or social benefit to the user? I welcome any responses to these comments. Please post.