Analyzing Execution Traces of Failed Programs for Materializing Chain of Infection

Debug is a software development task to find and correct the cause of a failed program execution. A diagnosis is a hypothesis to specify the cause and to explain the process for a program to fail.Evaluating the relevance of maintainers' internal thoughts is very important for conducting an experiment to evaluate a new debugger. We believe that analysis of such relevance gives us precise explanations of experimental results, which tells how a new debugger supports participants.We are now developing a dynamic analysis to obtain a data that enables us to evaluate the relevance of internal thoughts by a maintainer trying to make a diagnosis. Such a data, called materialized infection, enables us to check if execution points referenced by a maintainer implement the chain of infection to make a program fails. In this paper we explain the concepts of our dynamic analysis and discuss its limitations and future work.