A multi-tiered classification scheme for component retrieval

Software reuse demands that existing components must be readily incorporated into new products. These components need to be suitably classified to enable efficient retrieval, and if necessary, adaption. Typically, four schemes have been previously employed: Tree Text, Enumerated, Attribute Value and Faceted classification. The individual classification schemes are unable to solve the problems associated with component classification. As new techniques are developed for storing and presenting information, new classification schemes are required that utilise these technologies effectively. Multimedia for example, provides the developer with a number of viewpoints for documenting the component semantics. The paper looks at each of the classification techniques above, and proposes a new method for classifying the component details within a repository. Three classification strata are employed for the domain/setting, the functional and the presentational aspects of the component