Problem-solution mapping in object-oriented design

Six expert Smalltalk programmers and three expert procedural programmers were observed as they worked on a gourmet shopping design problem; they were asked to think aloud about what was going through their minds as they worked. These verbal protocols were recorded and examined for ways in which the programmers' understanding of the problem domain affected the design process; most of our examples are from the three Smalltalk programmers who focussed most on the mapping from problem to solution. We characterize the problem entities that did appear as solution objects, the active nature of the mapping process, and ways in which the resultant objects went beyond their problem analogs.