On analytic maintenance process data at the global and the detailed levels: a case study

Concerns the empirical analysis of data collected during the maintenance phase of a large operating system. The data concern notices that are issued to the commercial users of the system, detailing how to avoid, circumvent or patch the error that it describes until the latter can be properly corrected in a subsequent official release of the operating system. Notices had been collected, compiled, and statistically analyzed for more than three years, first at the global level to determine whether they conformed to previous observations, and then at a sublevel, that of each successive release for the article, in an attempt to determine whether patterns exist at sublevels, how such patterns determine those, if any, at the global level, and whether they can be utilized for the creation of models to predict the behaviors of individual components of the system. The analysis shows that maintenance data observed at the global level tend to show few, if any, patterns. It appears that the number of components and the concealment of their individual patterns, and hence the strength of the global invariance and dynamics, are functions of the system's complexity.<
>