Replication of the First Controlled Experiment on the Usefulness of Design Patterns: Detailed Description and Evaluation

  • Name:

    Technical Article

  • Author:

    Douglas C. Schmidt
    Lutz Prechelt
    Barbara Unger

  • Summary

    Advocates of software design patterns claim that using design patterns improves communication between software developers. The controlled experiment that we describe in this report tests the hypotheses that software maintainers of well-structured, well-documented software containing design patterns can make changes (1) faster and (2) with less errors if the use of patterns is explicitly documented in the software. The experiment was performed with 22 participants of a university course on C++ and design patterns. It is similar to a previous experiment performed in Karlsruhe. For one of the two experiment tasks the experiment finds that both hypotheses appear to be true. For the other task the results are inconclusive, presumably because the task was too difficult for the given experience level of the subjects.

    Bibtex

    @techreport{,
    author={Douglas C. Schmidt, Lutz Prechelt, Barbara Unger},
    title={Replication of the First Controlled Experiment on the Usefulness of Design Patterns: Detailed Description and Evaluation},
    year=1997,
    month=Dec,
    url={http://ps.ipd.kit.edu/downloads/ta_1997_replication_usefulness_design_patterns.pdf},
    abstract={Advocates of software design patterns claim that using design patterns improves communication between software developers. The controlled experiment that we describe in this report tests the hypotheses that software maintainers of well-structured, well-documented software containing design patterns can make changes (1) faster and (2) with less errors if the use of patterns is explicitly documented in the software. The experiment was performed with 22 participants of a university course on C++ and design patterns. It is similar to a previous experiment performed in Karlsruhe. For one of the two experiment tasks the experiment finds that both hypotheses appear to be true. For the other task the results are inconclusive, presumably because the task was too difficult for the given experience level of the subjects.},
    organization={Washington University, St. Louis, MO, Deptartment of Coputer Science},
    pages={77},
    }
  • In:

    Washington University, St. Louis, MO, Deptartment of Coputer Science 

  • Year:

    1997

  • Links:
Associated Project
Title