On the Ability of Functional Size Measurement Methods to Size Complex Software Applications

The most popular Functional Size Measurement methods, namely IFPUG Function Point Analysis and the COSMIC method, adopt a concept of “functionality” that is based mainly on the data involved in functions and data movements. Neither of the mentioned methods takes directly into consideration the amount of data processing involved in a process. Functional size measures are often used as a basis for estimating the effort required for software development, and it is known that development effort does depend on the amount of data processing code to be written. Thus, it is interesting to investigate to what extent the most popular functional size measures represent the functional processing features of requirements and, consequently, the amount of data processing code to be written. To this end, we consider a few applications that provide similar functionality, but require different amounts of data processing. These applications are then measured via both functional size measurement methods and traditional size measures (such as Lines of Code). A comparison of the obtained measures shows that differences among the applications are best represented by differences in Lines of Code. It is likely that the actual size of an application that requires substantial amounts of data processing is not fully represented by functional size measures. In summary, the paper shows that not taking into account data processing dramatically limits the expressiveness of the size measures. Practitioners that use size measures for effort estimation should complement functional size measures with measures that quantify data processing, to get precise effort estimates