The purpose of this case study was to document and evaluate the application of a top-down design methodology (Foley & van Dam, 1982) to a pre-existing computer system to test the methodology's usefulness as well as to gain insights into the design process itself. System experts advocate design of a system “top-down” instead of “bottom-up” as a way to sequentially examine the complex task of interface design while allowing re-examination of previous steps in that design (Foley, 1981). The study involved a menu-based, mini-computer system designed at Goddard Space Flight Center called the Mission Planning Terminal (MPT). The MPT will be used at Goddard for planning and scheduling of satellite activities through the NASA Network Control Center (NCC). The scheduler/analyst's task includes submitting a schedule of activities for his mission, transmitting it to NCC, and then modifying the returned schedule, if necessary, using the MPT. The top-down design process is distinctly divided into four phases: conceptual, semantic, syntactic, and lexical (Foley, 1981). The first phase, conceptual, consists of defining key application concepts needed by the user. The semantic phases involve defining meanings such as information needed in order to use an object. The syntactic design defines sequences of inputs (similar to English grammar rules) and outputs (the two and three dimensional organization of the display). The last step, lexical design, describes how words in the input/output sequence are formed from the existing hardware input (Foley & van Dam, 1982). The top-down methodology was applied using MPT documentation and interviews with the designers. During this process, it became clear that although a conceptual model of the MPT existed somewhere, it was never recorded. This led to numerous attempts to extract the main conceptual components of the system from the software operations documents which were constantly changed and were often incomplete. Finally, based on preliminary screen designs, state diagrams were constructed to map out components of the system. By characterization of the MPT in this way (using state diagrams), a clearer picture emerged that finally led to understanding the conceptual model. Once the conceptual model was extracted, redesign of the system, using the top-down method, quickly followed. This case study clearly emphasizes the need for a complete and accurate conceptual model if a top-down approach is to be applied. When redesigning an existing system, it frequently becomes necessary to “extract” this model in a bottom-up manner as was the case here.
A Unified Conceptual Model for Data Warehouses
