Increasing demand for quality software that meets the criteria of being economical, portable, efficient, reliable, robust, and upgradable has been met with the component based approach to software development. A further constraint today is the availability of trained man-power pool. A component is used on either an “as-is” basis (adoption) or on a “parameterized” basis (adaption). New components are added to the pool when necessary. The term “template” is frequently used for components. Component Based Software Engineering (CBSE) has thus emerged as the new effective paradigm in the 90’s. Measurement techniques in CBSE have, however, been rather limited. Even the so called ‘global’ methods like Mark II Function Points are not quite suitable when it comes to CBSE. The paper reviews the research work of various authors in the area of metrics in CBSE. Many of these authors have focused on extending Object-Oriented metrics like C-K metrics. Three areas under close scrutiny are complexity, customizability, and reusability. These authors have based their approaches more on theoretical aspects, while empirical studies have been neglected. Even the Verner- Tate empirical studies are based on the Fourth Generation Language Paradigm of the 80’s (where traditional metrics can be applied with relative ease). Today we would not accept them as CBSE. The authors’ suggestion is to include architectural styles, architectural patterns, and frameworks while studying metrics. Two examples are OMG’s CORBA and Microsoft’s DCOM. A number of frameworks are available in the Open Source community. The paper presents the findings of a small study of projects undertaken by students in the web publishing area. It is the authors’ contention that metrics must be specific to the framework used for implementation. It is a relatively simple task to mount an instrument in a framework such as TYPO3 (Open Source Web Content Management System Framework) for measurement purposes. This is illustrated in our paper. It would take a number of such studies to validate various metric models proposed by authors in the area of CBSE.
