Theoretical and Probabilistic Conversion Model between FPA and COSMIC Measurement Method

Abstract

Functional Size Measurement (FSM) has been widely adopted by software organizations over the past years. In particular, Function Point Analysis (FPA) measurement method as the eldest and the most globally FSM method used. In the middle of 90s, COSMIC measurement method was designed to cover the limitations of previous FSM methods, and to size wider scope of application domains. COSMIC measurement method had received special attention due to its novelties and merits it offers. In this respect, many organizations have moved from using FPA to the COSMIC measurement method, mainly because FPA cannot size as many software project types as COSMIC. Since there have been many projects sized with FPA are available in various databases, researchers undertook to compare the Unadjusted FPA (UFP) size and COSMIC Function Point (CFP) size unit statistically, towards converting FPA data in to its corresponding value in CFP. In the literature, there is only one conversion model that converts the size at the level of Basic Functional Components (BFCs) between the two methods. This paper examines the accuracy of the conversion model that converts the size at the level of BFCs. Moreover, this paper presents a theoretical and probabilistic conversion model that converts the FPA transaction functions size to the most likely corresponding CFP size at the level of BFCs. The proposed model considers the primary intent and the processing logic forms of each FPA transaction function BFC, measurement best practices and COSMIC method rules. The proposed model has been validated using 14 industrial business applications. The model was found to adequately and high accurately convert all the projects tested precisely. The model might be useful for companies that are using both methods or planning to adopt the COSMIC method as a conversion or validation procedure.