Review

The presentation notes for an early version of this paper inspired my PhD thesis. In those notes, Ray Reiter showed how to formalize a relational database as a set of logic formulas, and showed how queries also have a clean logic semantics. He talked about how one could represent null values in logic, and briefly mentioned how null values made it really hard, or perhaps impossible, to do updates. I thought, "I know how to do those updates!", and off I went to do my PhD research. The notes for the paper, and eventually the paper itself, were classic Reiter: not an extra or unnecessary word, and everything as clean and tidy as a reader could desire.