An empirical study of function points analysis reliability
Michael J. Connolley · DSpace@MIT (Massachusetts Institute of Technology) · 1990
The Function Points methodology was developed by Allan Albrecht to help measure the size of software programs. The methodology quantifies the amount of functional-ity provided by an application and expresses it in a metric called the Function Point. Such size measurements are needed to quantify development and maintenance productivity, and to estimate development work effort. The Function Points method-ology has been adopted by over 500 major corporations and is rapidly growing in popularity. Numerous variants of the Function Points methodology have been proposed. Despite the growing use of Function Points, these is a widespread perception that the Function Points metric is not reliable; that different people estimating the same application will get much different results. This study looks at the reliability of the Function Points metric in two ways: first it quantifies the reliability of Function Point counts made by raters using exactly the same methodology; and second, it quantifies the reliability of counts made using two