Author List: Wang, Shouhong;
Journal of Management Information Systems, 1996, Volume 12, Issue 4, Page 117-141.
Object-oriented approaches have received attention in management information systems development due to the advantages over the traditional approaches claimed by the proponents of the object-oriented approaches. To describe how people actually perform object-oriented analyses, this paper formalizes an object-oriented systems analysis approach. Protocol analyses of seven systems analysts, who were experienced with structured analysis, were used to formalize procedures for object-oriented analysis. The protocol analyses revealed four fundamental types of object classes in a management information system. They are: input, output, physiomorphic, and event object classes. The identification of input, output, and event classes depends upon the problem domain being analyzed. Physiomorphic classes, on the other hand, are more likely to match a schema that is perceived by the analyst based on his a priori knowledge about the problem domain classes. The protocol analyses also revealed that object classes are identified in an ad hoc manner; however, when checking the analysis, depth-first or breadth-first searching methods are often used. These, in turn, are controlled by a global backward or forward tracing strategy. Tracing the origin of data in messages was the strategy used for checking the completeness of the analysis. An experiment was conducted to compare the protocol-based object-oriented method and structured analysis. Thirty-two students who had no previous systems analysis experience were trained and then completed a problem using both techniques. The protocol-based method produced analyses that more closely matched the problem. Furthermore, it required less time to complete an analysis, and it was perceived as easier to use by the participants than the structured analysis method.
Keywords: management information systems analysis; object-oriented systems analysis; protocol analysis.
Algorithm:

List of Topics

#60 0.371 analysis techniques structured categories protocol used evolution support methods protocols verbal improve object-oriented difficulties analyses category benchmark comparison provided recognition
#216 0.163 conceptual model modeling object-oriented domain models entities representation understanding diagrams schema semantic attributes represented representing object relationships concepts classes entity-relationship
#240 0.123 systems information management development presented function article discussed model personnel general organization described presents finally computer-based role examined functional components
#147 0.112 process problem method technique experts using formation identification implicit analysis common proactive input improvements identify traditional stages identifying explicit setting
#9 0.080 using subjects results study experiment did conducted task time used experienced use preference experimental presented decision-making empirical significantly effects better
#10 0.073 strategies strategy based effort paper different findings approach suggest useful choice specific attributes explain effective affect employ particular online control