function point analysis

views updated

function point analysis A method derived originally by Albrecht at IBM to estimate the relative complexity and work content of developing a software system. The requirements of a software system are analyzed for five categories: inputs, outputs, files, interfaces, and enquiries. Each of these is then classified into parts that are simple, average, or complex. The results are represented as counts in a matrix with a total of 15 cells; each raw count is weighted by multiplying it by standard factors, and the unadjusted function point count, U, is then obtained by summing each cell-weighted value.

The processing complexity for the software is estimated for each of 14 general characteristics that cover the type of product, and how it is to be used and installed. For each characteristic a value is selected to represent its scale of influence. The 14 values are summed to give the processing complexity adjustment, PC, which will range from 0 to 70. The PC is used to calculate the adjusted function point score from its unadjusted score U. A measure of the work involved in developing the software is then obtained from a formula that allows for further score adjustments where necessary.