Could micro patterns be used as software stability indicator?

Micro patterns can be a useful proxy for the quality of software. Classes matching certain categories of micro patterns were shown to be more fault prone than others, and those classes that do not correspond to any category of micro patterns were shown to be more likely to be faulty. In this paper we present a preliminary study of traditional software metrics and micro patterns in three versions of Eclipse (2.1, 3.0, 3.1) in order to understand if it is possible to relate the stability of a software system with micro patterns.