Information presentation through default displays
Sakunthala Gnanamgari · Scholarly Commons (University of Pennsylvania) · 1981
Methods of presenting information automatically are, by and large, limited to variations of tabular form. Reading a sequence of lines and understanding their import is a tedious job reminding people of the old adage, 'A picture is worth a thousand words.' As a result, efforts are now being directed toward presenting such data graphically. Unfortunately, using graphic devices can be a complex process, requiring days or even weeks of training in the area of Computer Graphics. Until recently it has been difficult for a naive user to create a graphic display of information. The long range goal of this research is to have an intelligent system which helps users in the graphical display of data by performing the task of a graphic artist. One immediate problem is that there is a gap between the way a user conceives of a graphic display and the way the machine does. Another problem is that often a user will not think to make explicit what s/he does not care about or what s/he believes the system already knows or is able to infer. What is needed is a graphic-expert system that, on the one hand, is at an appropriate conceptual level to state those things that the user cares about and, on the other, provides appropriate defaults to take care of what the user considers to be trivial or obvious. The immediate objective, accomplished in this research, involves the automatic selection of graphic display formats. This selection is based on the formally characterizable properties of the data to be displayed, the user's particular objectives, principles of graphic design and the results of studies in the area of human factors and effective displays. The major contributions of this research lie in solving the joint problems of (1)deducing, in lieu of explicit specification from a user, an effective graphic presentation of tabular data, and (2)identifying design factors that are necessary to decide upon an effective display format for the given information. The contributions in these two areas derive directly from the method that is suggested in this thesis of formalizing the way that tabular data is viewed. The most significant point of this work is the new direction toward providing automatic graphic displays. Until now, the emphasis has been on enhancing systems for automatic textual display so that they are easily usable by non-programmers and can provide defaults including, for example, automatic paging and right justified margins. This research has established the groundwork for developing automatic graphic display systems for naive users by providing the same or greater degree of default capabilities and ease of use.