Measuring complexity and coverage of software specifications

    Authors

    G. Walton;J. H. Poore

    Abbreviated Journal Title

    Inf. Softw. Technol.

    Keywords

    Markov chain; usage model; statistical testing; path coverage; complexity; optimization; Computer Science, Information Systems; Computer Science, Software; Engineering

    Abstract

    Coverage testing in the context of Markov chain usage models refers to coverage of a model of the specification and profile of intended use, rather than coverage of the code that implements the specification. A new measure of specification complexity based on the number of statistically typical paths through the model of the specification is derived. Formulae are presented to compute bounds on the expected number of test cases required to achieve state and are coverage. Formulae are presented to compare different usage models with respect to the amount of testing required to achieve coverage of typical paths. Convexity properties are established for these formulae to facilitate their use in optimization calculations that are used to generate transition probabilities for the usage models. (C) 2000 Elsevier Science B.V. All rights reserved.

    Journal Title

    Information and Software Technology

    Volume

    42

    Issue/Number

    12

    Publication Date

    1-1-2000

    Document Type

    Article

    Language

    English

    First Page

    859

    Last Page

    872

    WOS Identifier

    WOS:000089543000009

    ISSN

    0950-5849

    Share

    COinS