Reconcilable differences (Q639856)

The paper analyzes several relevant problems in database theory, view maintenance and view adaptation, by looking at them in a unified fashion as instances of the problem of query reformulation using views (where relational difference must be allowed). The authors introduce the novel \(\mathbb{Z}\)-semantics for relational databases and show that, under this semantics, queries have a normal form expressed as the difference of two otherwise positive expression queries. The existence of this specific form implies the decidability of the equivalence problem within the so obtained formal system. A sound and complete algorithm for query reformulation using views over \(\mathbb{Z}\)-relations is then provided. Finally, the authors relate their framework to more standard forms of relational semantics, namely, the set and the bag ones. I would like to state that I found this paper a delightful reading, for two main reasons. First of all, I greatly appreciated the basic idea underlying \(\mathbb{Z}\)-semantics: so simple and clear; this is one of those cases when a reader asks himself, how come no one has proposed this before? In science, such kinds of ideas often turn out to be the most successful ones in the long run. Secondly, although the technical material treated in the paper is far from being straightforward, reading the paper is very pleasant and, I would dare to say, easy, even for the non-expert reader. Why is this? The driving forces here appear to be a very polished prose and a well-designed presentation ``plot''. Summarizing, a very nice paper which, I dare to affirm, will probably become a ``classic'' in database theory.