From critiquing to representational talkback: Computer support for revealing features in design

Computational critiquing mechanisms support designers in refining a partial design solution in an alternating cycle of reflection and action. This paper argues that critiquing mechanisms can support designers not only in refining a partial solution, but also in gaining a better understanding of the problem. We further argue that different types of critiquing are possible, ranging from conventional rule-based messages to an implicit type of critiquing based on the notion of "representational talkback" - representations that can reveal to the user otherwise implicit features of a design. We support these claims with user studies of three types of design support systems: KID, IAM-EMMA, and ART.