The spreadsheet is one of the most important decision support tools used by businesses. It has the reputation of being powerful, versatile, and user-friendly, so the practice of building a spreadsheet model has not changed. Lest we be complacent, the authors caution us that spreadsheet design errors are prevalent and difficult to detect, especially if one just looks at the spreadsheet’s physical appearance, which can be deceiving. What appears on the spreadsheet may not have anything to do with the logical structure it represents. This should be a scary thought for businesses and the customers they serve.

Since a theory of spreadsheet analysis and design has yet to emerge, the concern of the authors is to provide the first step to endowing the conventional spreadsheet program with the functionality of a formal spreadsheet model management system, thus coming up with a business-wide spreadsheet model management system. They hope that such a system would produce a well-designed, well-documented spreadsheet that can communicate with other models and data resources in the organization, so that one could move data from one platform to another, as in current database management systems, and have the ability to change assumptions and inspect their effect.

The authors try to show that there are three approaches that could be used: top-down, inside-out, and bottom-up. They prefer the bottom-up approach because they believe in using data sets from existing spreadsheets instead of building new models. They consider the interplay of the physical and logical views of spreadsheet models to contain the following elements: schema, data, editorial, and binding. They hope to build relational spreadsheets in which a change in one element will affect another. The user can then build spreadsheet models in his or her favorite spreadsheet program without having to learn a new modeling language.

I enjoyed the challenge of reading this paper. Although at times it has a tendency to be technical, and there are empty tables, the lay reader should be able to follow the logic of what the authors are trying to say. The paper raises issues that I am sure businesses have also considered regarding their use of the spreadsheet as a management tool. The compatibility of spreadsheets used by different people for the same purposes can only be achieved by copying an original, erasing the contents, and then using it. But once the original is changed, the changes are not reflected in the copies, so the possibilities for errors are great. Imagine the time that can be saved by making spreadsheets relational. I commend the authors for their efforts. I am sure that relational spreadsheets will be forthcoming as a result of their seminal work.