Constraint-based hypertext for argumentation

  • Authors:

  • Paul Smolensky;Brigham Bell;Barbara Fox;Roger King;Clayton Lewis

  • Affiliations:

  • Department of Computer Science and Institute of Cognitive Science, University of Colorado, Boulder, CO;Department of Computer Science, University of Colorado, Boulder, CO;Department of Linguistics and Institute of Cognitive Science, University of Colorado, Boulder, CO;Department of Computer Science, University of Colorado, Boulder, CO;Department of Computer Science and Institute of Cognitive Science, University of Colorado, Boulder, CO

  • Venue:
  • HYPERTEXT '87 Proceedings of the ACM conference on Hypertext
  • Year:
  • 1987

Quantified Score

Hi-index 0.00

Visualization

Abstract

In this paper we describe a hypertext system we are developing for the support of reasoned argumentation: the EUCLID project. We use the project to address two general problems arising with hypertext: the problems of controlling user/document interaction, and the problem of controlling the screen. We suggest that guiding users' interaction with hypertext is difficult because of the unique form of discourse that hypertext represents, and that structuring user/document interaction can be achieved through specializing to a particular type of material and designing the hypertext system to respect the particular discourse structure characteristic of that material. EUCLID's design is tuned to the structure of reasoned discourse. The problem of screen management in EUCLID is a serious one, because our presentation of complex arguments requires mapping the complex logical relations between parts of realistic arguments onto complex spatial relations between items in the display. We describe a general system we are developing which provides this high degree of control for hypertext screen management. This system represents a constraint-based approach to hypertext, in which the items from the underlying database that are to be displayed may each contribute a number of constraints on the layout; a general constraint-satisfier then computes a screen layout that simultaneously satisfies these constraints. Each time an item is to be added to or deleted from the screen, the constraint set is adjusted and the screen layout is recomputed; thus the spatial relationships on the screen provide at all times a veridical representation of the underlying relations between displayed database items. This kind of strong screen control is demanded by hypertext applications which, like ours, are fine grained: the number of nodes and links being displayed number in the hundreds.