• DocumentCode
    2190877
  • Title

    Possible implications of design decisions based on predictions

  • Author

    Larsson, Magnus ; Crnkovic, Ivica

  • Author_Institution
    ABB Corp. Res., Vasteras
  • fYear
    2004
  • fDate
    7-10 June 2004
  • Firstpage
    651
  • Abstract
    Software systems and applications are increasingly constructed as assemblies of preexisting components. This makes software development cheaper and faster, and results in more favorable preconditions for achieving higher quality. This approach, however, introduces several problems, most of them originating from the fact that preexisting software components behave as black boxes. One problem is that it is difficult to analyze the properties of systems in which they are incorporated. To simplify the evaluation of system properties, different techniques have been developed to predict the behavior of systems on the basis of the properties of the constituent components. Because many cannot be formally specified, these techniques make use of statistical terms such as probability or mean value to express system properties. This paper discusses ethical aspects of the interpretation of such predictions. This problem is characteristic of many domains (data mining, safety-critical systems, etc.) but it is inherent in component-based software development
  • Keywords
    data mining; decision making; ethical aspects; formal specification; object-oriented programming; probability; risk analysis; software development management; software quality; black boxes; component-based software development; data mining; design decisions; ethical aspects; probability; safety-critical systems; software components; software development; software systems; Application software; Assembly systems; Companies; Computer science; Ethics; Probability; Programming; Software quality; Software systems; Testing;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Information Technology Interfaces, 2004. 26th International Conference on
  • Conference_Location
    Cavtat
  • Print_ISBN
    953-96769-9-1
  • Type

    conf

  • Filename
    1372495