software measurement methods an analysis of two designs软件度量方法的分析两个设计.pdfVIP

Journal of Software Engineering and Applications, 2012, 5, 797-809 797 /10.4236/jsea.2012.510092 Published Online October 2012 (http://www.SciRP.org/journal/jsea) Software Measurement Methods: An Analysis of Two Designs Jean-Marc Desharnais, Alain Abran Department of Software Engineering and IT, École de Technologie Supérieure, Montreal, Canada. Email: desharnaisjm@, alain.abran@etsmtl.ca Received August 22nd, 2012; revised September 24th, 2012; accepted October 6th, 2012 ABSTRACT In software engineering, software measures are often proposed without precise identification of the measurable concepts they attempt to quantify: consequently, the numbers obtained are challenging to reproduce in different measurement contexts and to interpret, either as base measures or in combination as derived measures. The lack of consistency when using base measures in data collection can affect both data preparation and data analysis. This paper analyzes the simi- larities and differences across three different views of measurement methods (ISO International Vocabulary on Metro- logy, ISO 15939, and ISO 25021), and uses a process proposed for the design of software measurement methods to analyze two examples of such methods selected from the literature. Keywords: Software Measures; Base Measures; Derived Measures; Measurement Method; Attributes; Software Quality Model; Metrology; Software Metrics 1. Introduction quantity to be measured (measurand1), the measurement procedure, the measurement results or models of rela- In the sciences and in engineering, the consensus on rigo- tionships across multiple measures, and measurement of rous measu