Please use this identifier to cite or link to this item: http://hdl.handle.net/10525/730

 Title: A Sensitive Metric of Class Cohesion Authors: Fernández, LuisPeña, Rosalía Keywords: Object-Oriented ProgrammingMetrics MeasurementQuality Analysis and Evaluation Issue Date: 2006 Publisher: Institute of Information Theories and Applications FOI ITHEA Abstract: Metrics estimate the quality of different aspects of software. In particular, cohesion indicates how well the parts of a system hold together. A metric to evaluate class cohesion is important in object-oriented programming because it gives an indication of a good design of classes. There are several proposals of metrics for class cohesion but they have several problems (for instance, low discrimination). In this paper, a new metric to evaluate class cohesion is proposed, called SCOM, which has several relevant features. It has an intuitive and analytical formulation, what is necessary to apply it to large-size software systems. It is normalized to produce values in the range [0..1], thus yielding meaningful values. It is also more sensitive than those previously reported in the literature. The attributes and methods used to evaluate SCOM are unambiguously stated. SCOM has an analytical threshold, which is a very useful but rare feature in software metrics. We assess the metric with several sample cases, showing that it gives more sensitive values than other well know cohesion metrics. URI: http://hdl.handle.net/10525/730 ISSN: 1313-0463 Appears in Collections: Volume 13 Number 1

Files in This Item:

File Description SizeFormat