The limits of empirical studies of software engineering

Some advocates of empirical studies of software engineeringappear to be claiming that empirical studies alone cantellus how we should do software development.This paperargues that what can be learned from empirical studies,while important, is very limited.Mathematical studies andempirical studies must be seen as mutually supportiveways to increase our knowledge of software developmentmethods.The role of empirical studies should be limited toconfirming that what works in theory can actually be used(and useful) in practice.