On the danger of developing measures without clarifying concepts [software measurement]

Abstract

Defining concept before developing theory must be a fundamental requirement of any science. We show that this requirement, though seemingly trivial, is to a large extent not respected in software measurement research. Many basic software quality concepts are used ambiguously without a clear definition and consequently their measures, though growing in number, have no serious scientific foundation and risk being misleading in further research. We demonstrate the differences between concept and measure, show the risks to develop measures without clarifying concepts, and underline the paramount importance of concept clarification and definition for the development of software measures.