Defects in object-oriented designs are more costly to repair if they are detected later in the software development process. Early detection of defects should significantly reduce development costs. Apart from detecting obvious syntactic defects of the model, a major challenge is to check whether the requirements model actually corresponds to the reality it is supposed to model.

The goal of this book chapter is to present a method, called UML-HAZOP, that would systematically generate checklists for unified modeling language (UML) models. Derived from the hazard and operability (HAZOP) method for safety-critical systems, this method follows all of the connections in a model (currently associations and generalizations in class diagrams), and, based on their attributes, suggests all of the possible defects. A human analyst then has to decide whether each one is really a defect. While scrutinizing a model in this way helps to detect conceptual defects, the human effort involved appears to pay off only for safety-critical systems.

Unfortunately, the method is presented only briefly. No detailed description or example is given, and the reader is referred to an earlier journal paper by the same authors. While the chapter is already short, a large part of it is spent on unimportant technical details about the tool implementing the method. Even though the abstract is promising, the chapter itself, in my opinion, is not interesting for any audience; a reader interested in the method has no other choice than to look up the journal paper [1].