Probing methods for automatic error resolution in a heterogeneous software environment

We investigate the feasibility to improve the error resolution through automation like natural language-based and statistical analysis algorithms, in order to detect errors and security issues, and to convert error messages from encrypted into human-readable ones in a heterogonous software environment. To reach this goal we study a real case using the data extracted from PopCon, a package used for the population of CMS Condition Databases, that is embedded into CMS Software framework, CMSSW, and relies on different underlying applications such as ORACLE, POOL, CORAL in order to perform database transactions.