Toward precise measurements using software normalization

There has been a lot of interest in defining appropriate ways of measuring the attributes of software. We think these measures should accurately represent those attributes which they quantify. However, these software measures are usually not precise enough because they measure elements that need not be included. Reducing redundant structures of the program before they are measured is a simple yet important concept. We propose a strategy of software normalization for enhancing software measurement. Most importantly, this program normalization allows those metrics to measure what is meant to be measured and avoid imprecise results.