Preserving knowledge in design projects: what designers need to know

Reviewer: Tom DeMarco

Herbsleb and Kuwana have performed an empirical study of information types needed by software designers. They analyzed three projects, two at US companies and one in Japan. Though design rationale data were collected in different ways on the three projects, it was possible to categorize them on a common taxonomy. Subsequent instances of information “fetch” by developers were then intercepted and tallied by taxonomic type. The results were surprisingly similar across the projects. Capture of design rationale information is desirable in the abstract, but when you come face to face with the specifics of capturing it on your project, the quantities of data can be daunting. Herbsleb and Kuwana's findings give us a way to pick and choose among design rationale items and decide sensibly which items could justify the expense of formal capture. The methodology used in the paper is interesting in itself. It provides a scheme for typing design rationale information. The different classes of such information identified by the authors could form a useful checklist to guide new managers, or an organizational scheme for project documents.