Engenharia de confiabilidade de software: um mapeamento sistemático

Abstract

Computer systems are present in practically all of the main processes of modern society. This high dependency means that failures in such systems have a direct impact upon people lives. Computer system failures are mainly caused by software defects. The software reliability engineering (SRE) is the field that studies how to make software more reliable. SRE is a multi-discipline field, which integrates in essence, probability, statistics, and computer science. The research areas in SRE are varied, so, diversity questions could be raised as to: Which fields of SRE have been most recently investigated? Which are the international SRE groups most active? Which problems are covered by the most cited SRE studies? Which are the emerging topics in SRE? Therefore, this study presents the results from a systematic mapping study on software reliability. This study considered the last 12 years of publications covering 92 publication venues (journals and international conferences), with a total sampling of 300 papers. Statistical techniques were used for clustering papers in terms of similarity, as well as for identifying the research topics of software reliability. The results showed that studies which adopt a purely theoretical approach were predominantly greater in number than empirical studies. However, we observe that over the last few years the theoretical-empirical approach has become more relevant. Besides this, it was also observed that the factor external collaboration presented itself as one of the most evident in those research centers of greater influence. Finally, among the research topics covered in this study, we identified a trend through the results, towards a demand in theoretical-empirical research in software testing for the coming years.