Software Engineering Measurement [chapter]

2014 Chapman & Hall/CRC Innovations in Software Engineering and Software Development Series  
An approach that employs quantitative measurements makes it easier to understand, evaluate, predict and monitor the development process and the software products developed. This need for measurement was pointed out by Basili (1995) , who notes that the measurement process is an excellent conceptual mechanism for learning what works and what doesn't. The quantitative approach is, in fact, a very important component of the management process in most departments of an organization: market
more » ... selection, design and development of new products and services, production, delivery, billing, finance, even human resources. In all organizations, many of the quantitative components needed for decision making are found in the accounting systems, used not only for budgetary controls, but also for performance analyses and the taking of decisions that lead to all manner of changes within the organization. These accounting systems are, indeed, used in the overall process of production cost analysis; the same can be said of management engineering, where improvements to processes and products are made based on quantitative data collection systems. Software Metrics? But is this true of software management? What are the quantitative data collection systems for decision making? What quantitative models and measurements are available to Information Technology (IT) managers for analyzing productivity, assessing development and maintenance projects, and monitoring quality and productivity improvement programs? And how valid and reliable are the measurements and analytical models made available to IT managers? Finally, what approach should be used to analyze this validity and reliability? In IT, in both North America and Europe, quantitative measurements and models fall under what is traditionally known as "software metrics". Now, for many authors, including Fenton (1991), the English term "software metrics" encompasses a whole series of concepts related to a host of activities having common quantitative elements: -measurements and models for estimating cost and effort -productivity measurements and models -quality assurance and control -data collection -quality measurements and models -reliability models -performance evaluation -computational complexity -structure and complexity measurement systems. This term has not, however, been standardized, and the definitions of measurements and metrics vary depending on the author. The standardization work currently being done by ISO subcommittees in the field of software engineering to clarify the concepts and vocabulary and the analytical methods, must therefore be seen in this context. We would cite, for example, the ISO SC7-WG6 sub-committee, whose mandate is to develop quantitative quality measures for software products (ISO-9126), and the ISO SC7-WG12 sub-committee for the standardization of measures of the functional size of software (ISO 14143-1).
doi:10.1201/b17461-10 fatcat:nmfaoen3c5flrkhhg4utk4y6qm