Topic

Object-Oriented Product Metrics

Introduction

The field of OO product metrics is attracting many researchers’ interest, as metrics provide a mechanism for assessing product quality. Measurement is a key element of any engineering process. Many researchers have dedicated much of their time and effort to devising a single comprehensive measure of software complexity. According to Fenton [FEN94], such measure is an “impossible holy grail”.
Assuming Fenton was right and that there really isn’t a single comprehensive measure, a number of “independent indicators of the quality of analysis and design models” [PRE05, p.467] could be devised in order to measure various software attributes. There’s a great quantity of software tools for measuring and analysing a variety of product metrics. According to Pressman, “hundreds of metrics have been proposed for computer software, but not all provide practical support to the software engineer” [PRE05, p.469].

Journals, organizations, conferences

The various proposed metrics and others’ criticism of these are most usually published in the IEEE Transactions on Software Engineering (the most dominant source), Journal of Systems and Software, or IEEE Computer. Current research is presented in the Proceedings of the Symposium on Software Metrics (IEEE, published annually). There is a number of non-profit software metrics associations, such as UKSMA (United Kingdom Software Metrics Association), international organization ISBSG (International Software Benchmarking Standards Group), and individual research groups, for example the ESERG (Empirical Software Engineering Research Group of the Bournemouth University.
20th Annual UK Software Metrics Association Conference has taken place just a few weeks ago, on the 15th of December 2009. Metrikon 2009 – DASMA Metrik Kongress – at the Fraunhofer IESE will take place between 19th-20th November 2009 in Keiserslautern, Germany. There are workshops organised every year - International Workshop on Software Measurement (IWSM).

Leading researchers

The most prominent names in this area are McCabe, Henderson-Sellers, Chidamber, Kemerer, Fenton and Abreu. There are many other researchers and authors of great books on this topic, such as Stephen Kan (author of Metrics and Models in Software Quality Engineering).
McCabe being famous for his invention of cyclomatic complexity measure, Chidamber and Kemerer for their OO metrics suite, Henderson-Sellers and Fenton for their vast contributions to research in object-oriented methodologies, and Abreu for the MOOD (Metrics for Object Oriented Design) metrics suite.

History

History of OO Metrics naturally begins with the invention of the OO programming paradigm. “It was only in the 1980s that object-oriented design methods emerged. Object-oriented analysis methods emerged during the 1990s.” [GRA01]
The aforementioned Henderson-Sellers in cooperation with Edwards developed MOSES (1994), which was “the first OO method to include a full-blown development process, a metrics suite and an approach to reuse management” [GRA01].
McCabe’s devised the cyclomatic complexity measure in 1976, which was well before the OO programming paradigm became popular and modern OO languages (such as Java or C++) were created. Nevertheless, this measure is applicable to OO software design and is one of the most widely used measures, because it’s fundaments lie in sound mathematical theory (graph theory).
In the 90’s: Chidamber & Kemerer propose their OO metrics suite (1994). Cohesion metrics – LCOM - (how well the methods of a class are related to each other) are being formulated by many researchers. MOOD (1995) and MOOD2 (1998) metrics were devised by aforementioned Abreu.

Current trends

Currently the researchers are interested more in the areas of automated metric collection and analysis, visualisation of the collected metrics, and ways of applying the knowledge gained through metrics to improve the software ilities (i.e. maintainability, flexibility, extendibility, etc.) and overall product quality.

Problems with OO metrics

According to Pressman, measure that lead to a metric should be:

1. simple and computable
2. empirically and intuitively persuasive
3. consistent and objective
4. consistent in the use of units and dimensions
5. programming language independent
6. an effective mechanism for high-quality feedback

[PRE05, p.470]
Of course, each of these points poses a problem to the potential new useful metric. Sometimes the new metric simply requires too complex computation; other times its sole essence is so difficult to comprehend that no one will chose to use it. Some metrics are only applicable to a certain programming language, and some are just plain wrong, as they “violate the basic intuitive notions of what high-quality software really is” [PRE05, p.469].

[FEN94] Fenton, N., Software measurement: a necessary scientific basis. IEEE Transactions on Software Engineering, vol. SE-20, no. 3, March 1994, pp.199-206
[PRE05] Pressman, R., Software Engineering: A practitioner's approach, 6th edition. New York: McGraw-Hill, 2005
[GRA01] Graham, I., The history of object-oriented analysis and design methods. [On-line] Available from: http://uml-tutorials.trireme.com/uml_tutorial_1.htm Accessed on: 05-11-2009