11.2      Contributions to a Theory of Computer Support

The central contribution was to identify the key concept for a theory of computer support: interpretation. Although Winograd & Flores (1986), for instance, talked a lot about interpretation, they ranged across Heidegger’s (1927) framework and focused on its critique of technical rationality. Ironically, their proposed software example, the Coordinator program, suffered from a lack of respect for the importance of interpretative control by the users. They failed to take seriously the fact that there is no objective structure to a domain and that people should be supported in defining their own analyses, interpretations, and terminologies from their own perspectives. Support for interpretation is the ingredient missing from most traditional AI programs. This dissertation contributes the antidote: a recognition of the central role of interpretation and the impossibility of fully automating it. It is difficult to convey the potential importance of this contribution; that is why so many pages of the dissertation have been devoted to this theme.

The proposed theory of computer support is built squarely on the analysis of interpretation. This gives the theory a coherence and consistency missing from other theoretical frameworks in computer science (other than those based on strictly formal logical grounds). It demonstrates how philosophy (again, other than logic) can be put in the service of computer science.

Knowledge-based system design inevitably raises the question of the nature of knowledge. Some contributions have been made here. First, the varieties of knowledge or information have been categorized in terms of their origins in various phases of the process of interpretation. This includes not only tacit and explicit understanding, but also shared understanding and captured computer representations. Second, the idea of domain knowledge has been critiqued. Not only does knowledge in a design domain change as the related technologies and styles change and as the expertise of the field matures and grows, but every designer and every design team has their own domain knowledge. It is not simply that they each have different pieces of an underlying knowledge. Rather, to know is to know from a perspective, so there is no objective body of domain knowledge independent of what people know in their own ways, within their many perspectives. Third, the role of language in expressing knowledge has been emphasized. The emergence of interpersonal or operationalized knowledge from tacit experience takes place through discourse and assertion within situated interpretation. Correspondingly, an end-user language has an important role to play in computer support.

Go to top of this page

Return to Gerry Stahl's Home Page

Send email to Gerry.Stahl@drexel.edu

This page last modified on January 05, 2004