Infection size as a measure of bug severity

A simple bug in a program can influence a large part of the program execution by spreading throughout the state at runtime. This is known as program infection. The seriousness of bugs is usually measured by studying their external effects. However, such effects essentially derive from internal factors of a program. Our idea is to focus on internal factors, in particular the infection chain, to measure how serious a bug was. This allows reasoning about bugs from a new and potentially insightful perspective.