Efficient Specification-Based Component Retrieval

  • Authors:

  • John Penix;Perry Alexander

  • Affiliations:

  • NASA Ames Research Center, MS 269-3, Moffet Field, CA 94035 USA. jpenix@ptolemy.arc.nasa.gov;University of Cincinnati, Cincinnati, Ohio 45221-0030. perry.alexander@uc.edu

  • Venue:
  • Automated Software Engineering
  • Year:
  • 1999

Quantified Score

Hi-index 0.00

Visualization

Abstract

In this paper we present a mechanism for makingspecification-based component retrieval more efficient by limiting theamount of theorem proving required at query time. This is done byusing a classification scheme to reduce the number of specificationmatching proofs that are required to process a query. Components areclassified by assigning features that correspond to necessaryconditions implied by the component specifications. We show how thismethod of feature assignment can be used to approximate reusabilityrelationships between queries and library components. The set ofpossible classification features are formally defined, permittingautomation of the classification process. The classification processitself is made efficient by using a specialized theorem proving tacticto prove feature implication. The retrieval mechanism was implementedand evaluated experimentally using a library of list manipulationcomponents. The results indicate a better response time than existingformal approaches. The approach provides higher levels of consistencyand automation than informal methods, with comparable retrievalperformance.