ABSTRACT
More geographic information systems have failed through poor system design than from any other cause. Failure of the GIS is often not simple; failing systems frequently workto a degree-but they may meet only a portion of the critical user requirements, or they may not meet these requirements in a cost-and time-effective fashion. It is possible to move a hundred tons of sand over a distance of fifty miles in a small sports car, but what reasonable person would elect to do so if they were aware of the proper match between requirements and technology? The field of software engineering, a sub-discipline within computer science, has demonstrated that proper design of large, complex systems can take place and that the cost and time required for proper design can be recovered many times over. The application of these tools to the GIS area is overdue but it is starting to occur.
