Software component retrieval by composition.

Retrieving software component by composition helps user to build a new component using the existing components in the reuse component library when no single component in the component library can match the user requirements. This thesis proposes a method of retrieval-by-composition in which user specifies the composition structure and the system retrieves all possible compositions that fit into the structure specified by the user. In our method, semantic properties, including demand, transmission and length properties and used to reduce the composing and retrieval space. The semantic properties of internal components are inferred based on the user requirements. A prototype system called WISER is implemented, which helps user visually compose the composition structure, then does the inference on each composition component and displays all possible compositions. The implementation design thoughts and details are discussed with actual examples of retrieval-by-composition and many screen shots. The study shows that using semantic properties has greatly reduced the final search space in an inexpensive and effective way. Paper copy at Leddy Library: Theses & Major Papers - Basement, West Bldg. / Call Number: Thesis1997 .W8. Source: Masters Abstracts International, Volume: 37-02, page: 0639. Adviser: Young Gil Park. Thesis (M.Sc.)--University of Windsor (Canada), 1997