An interesting press release arrived in my Inbox the other day, from IFPUG, the International Function Point Users Group. IFPUG was formed in 1986 to promote the use of Function Point Analysis (FPA) as a tool for managing application software development. Briefly, Function Point Analysis quantifies the functions contained within software, relating directly to the business requirements that the software targets. To arrive at the actual function point measure, standardized criteria are applied to the business functions in the software, resulting in a numeric index of type and complexity for each function. The indices are totaled, and various factors that relate to the software as as whole are incorporated. The resulting number is the Function Point index, which measures the size and complexity of the software project.
IFPUG function points are now an ISO standard, and IFPUG is a prominent software measurement organization with a worldwide membership. IFPUG FPA is used increasingly as a basis for estimating software projects, defining requirements, tracking changes in scope, managing outsourcing contracts, and implementing process improvements. IFPUG offers a wealth of benefits with membership, including an annual conference, seminars and workshops, publications, and professional certification with the Certified Function Point Specialist (CFPS) program.
The group announced publication of an article, "The Statistically Unreliable Nature of Lines of Code," by Joe Schofield in the April 2005 Crosstalk, The Journal of Defense Software Engineering. The article challenges the use of the line-of-code measure as a reliable guideline for sizing software projects. Data for the article came from Personal Software ProcessSM courses taught at Sandia National Laboratories. Mary Bradley, IFPUG president, calls the article "an important addition to the software measurement body of knowledge. Anyone involved in measurement or estimation should read this article and understand the ramifications of using unreliable data."
Dianne Russell