The Classification of Representational Forms

Abstract

Knowledge access and ease of problem-solving, using technology or not, depends upon our choice of representation. Because of our unique facility with language and pictures, these two descriptions are often used to characterize most representational forms, or their combinations, such as flow charts, tables, trees, graphs, or lists. Such a characterization suggests that language and pictures are the principal underlying cognitive dimensions for representational forms. However, we show that when similarity-based scaling methods (multidimensional scaling, hierarchical clustering, and trajectory mapping) are used to relate user tasks that are supported by different representations, then a new categorization appears, namely, tables, trees, and procedures. This new arrangement of knowledge representations may aid interface designers in choosing an appropriate representation for their users' tasks.