Populating the software database

Abstract

The cost of software development could be reduced if relevant reusable software components could be retrieved efficiently. The few libraries currently in existence have no standard method for selecting components germane to the intended application. This thesis focuses on the actual formation and population of library components for an improved software library model proposed in Ref. 1. This library would provides the codes for users to implement the desired system in CAPS environment. The work reported here consists of: identifying candidate reusable components from the Booch Ada Library - by manually inspecting over 500 components; converting the components into a CAPS -compatible format based on the Prototyping System Description language (PSDL) via Ada-PSDL converter program; creating algebraic specifications to match the semantic description of each component manually; and manually organizing the library into a data structure based on the multi-level filtering concept. This work provides (1): the base and guidelines for the (a) criteria for a reusable component; (b) process of inspecting and importing components into CAPS reusable component library; (2): 75 reusable components to be released with CAPS 95 and used to test the user interface for retrieval via multi-level filtering. The process of populating reusable components is time intensive due to various manual processes. Inspecting and converting each component sometimes takes up to an hour for each. Current tools available can be rewritten i.e. the PSDL-Ada converter, to fully automate this process in accordance with the base and guidelines.