Case Study: COSMIC Approximate Sizing Approach Without Using Historical Data Francisco Valdés Souto École de Technologie Supérieure, University of Québec Dept. of Software Engineering Montréal, Canada francisco.valdes@spingere.com.mx Alain Abran École de Technologie Supérieure, University of Québec Dept. of Software Engineering Montréal, Canada alain.abran@etsmtl.ca Abstract—In mature engineering disciplines, international consensus can be reached on measurement, as evidenced through established measurement standards. In software engineering, there are 5 functional size measurement standards. These standards work best when the functionality to be measured is fully known, although this usually doesn’t happen in the early phases of software development. The techniques most often used to approximate the sizing of the software to be developed in the early phases involve historical data. However, gathering historical data is a challenge in itself. This paper proposes the use of a fuzzy logic model to approximate the functional size of a piece of software. Keywords—EPCU; COSMIC; Approximate Sizing; Fuzzy Logic; Functional Size; FSM. I. INTRODUCTION In mature engineering disciplines, there is international consensus on measurement, created through established measurement standards and their respective etalons 1 . In the software domain, quantitative international standards exist only for functional size measurement (FSM), as illustrated by the ISO 14143 series, in which key concepts of the entity and the attribute to be measured are prescribed. To date, the ISO has recognized five FSM methods for software as compliant with ISO 14143: • One of these is referred to as a 2 nd generation FSM method: COSMIC – ISO 19761. • Four are considered as 1 st generation FSM methods: MKII: ISO 20698, IFPUG: ISO 20926, NESMA: ISO 24570, and FISMA: ISO 29881. The recognized FSM methods work best when the information to be measured – the defined functionality (i.e. the functional requirements) – is fully known. This is most often 1 The Etalon, or Measurement Standard, is the realization of the definition of a given quantity, with a stated quantity value and associated measurement uncertainty, used as a reference [36]. not the case in the early phases of software development [22], when only the non detailed information is commonly found at this stage in the software life cycle [23]. Desharnais et al. point out that, due to a lack of good software documentation, it is not always possible to apply all the detailed rules, and measurers must fall back on approximation techniques for the subsets of requirements without enough details [20]. While the Desharnais et al. research work was related to the IFPUG method, it applies to any recognized functional size measurement methods for software. Desharnais et al. [20] identify a number of instances where the detailed measurement rules cannot be used: • The documentation is not precise enough for the application of the detailed measurement rules. • The amount of work required to apply the detailed measurement rules to obtain precise measures of the software, and the work required subsequently to update the measurement results, is perceived by management as being too expensive. • Qualified measurers are not available. Santillo [22] further states that the “functional size of software to be developed can be measured precisely [only] after the functional specification stage: this stage is often completed relatively late in the development process.” Consequently, a number of researchers have attempted to devise approximation approaches for measuring software functional size by analyzing data from completed projects. In 1992, Henderson et al. [26] investigated the relationship between function points and lines of code (the Backfiring method), but found wide variability in the results. In 1997, Meli [28] designed a technique which can produce two distinct types of size approximation: • Early Function Points (EFP), which is a faster version of the FP IFPUG 4.0 estimation method, and 2012 Joint Conference of the 22nd International Workshop on Software Measurement and the 2012 Seventh International Conference on Software Process and Product Measurement 978-0-7695-4840-1/12 $26.00 © 2012 IEEE DOI 10.1109/IWSM-MENSURA.2012.34 178