Experiencing Protocol analysis of computer-guided design tasks

In order to elaborate effective techniques and methods to conduct design tasks and to teach how to perform them, investigating the most effective cognitive approaches holds a paramount importance. Much of the effort devoted by the scientific community to gain a deeper understanding of the cognitive processes intervening during design tasks is addressed at eliciting the reasoning path through Protocol Analysis activities. The investigation of the most relevant issues related to the cognitive aspects through such a methodology consists in the interception of the reasonings followed by individuals and the characterization of such reflective dialogues through suitable taxonomies and frameworks. With the aim of acquiring the most complete comprehension about cognitive processes, the protocol analyses have substantially regarded design tasks carried out with the minimum extent of external boundaries affecting the thinking freedom. As a result, such experiences have extensively observed the behaviours of designers, assigned of a problem to be solved and working from scratch with traditional means (often paper and pen). The circumstances that are generally met when carrying out Protocol Analysis conflict with the growing diffusion of computer applications to support the design activities. The software systems and the media interface result indeed in the limitation of the designer possibilities during the task, causing an alteration of the followed cognitive process. The paper presents an investigation of such relevant issues aimed at providing useful insights about the usage of computer-based systems as viable means to perform Protocol Analysis tasks. According to these assumptions, Section 2 provides an overview about the documented employment of computers within the investigation of design protocols and explains the reasons for strenghtening its role in the field. Section 3 describes the application of a dialogue-based system, implemented in a computer framework and aimed at analyzing engineering problems for testing the possibility to perform the Protocol Analysis of a sample of generated design discourses. Section 4 discusses the results of such experiment from a methodological point of view and argues about further advantages with respect to traditional protocols. Finally, Section 5 is dedicated to draw the conclusions and illustrate the expected future developments, anticipating some basic requirements for the conception of software applications tailored to support the generation and the analysis of design protocols.