Computing Reviews
Today's Issue Hot Topics Search Browse Recommended My Account Log In
Review Help
Configuration knowledge of software product lines: a comprehensibility study
Cirilo E., Nunes I., Garcia A., de Lucena C.  VariComp 2011 (Proceedings of the 2nd International Workshop on Variability & Composition, Porto de Galinhas, Brazil, Mar 21, 2011)1-5.2011.Type:Proceedings
Date Reviewed: May 13 2011

We use the Latin square design for empirical experiments to isolate and quantify one or more process variables (or factors) in order to determine the effect the changes have on one or more response variables. This study--a 3x3 design--explores three techniques for applying configuration knowledge: annotation, general-purpose modeling, and domain-specific modeling. The two co-factors (nuisance variables) are subject expertise and product line configuration items.

The study had two objectives: to isolate which technique is most powerful with regard to configuration knowledge, as measured by the number of correct answers to a questionnaire and by the amount of time needed to complete the questionnaire, and to isolate whether or not individual subject expertise is a significant factor. The study results cannot be considered conclusive.

A couple of factors threaten the internal validity of the study. For example, a Latin square 3x3 design should be executed for nine runs; this experiment executed two runs. Also, there were only six subjects in the study (each subject took a 25-question questionnaire). A subject pool of six is not sufficient to isolate the main effect or determine the degree of interaction effects.

The researchers asked the subjects to self-report their levels of expertise. One subject self-rated as highest in the group, yet scored the lowest; another subject self-rated as almost the lowest of the group, yet scored almost four times higher than the lowest. The researchers did not apply a control group.

The questionnaire used to evaluate the subjects’ knowledge was not validated as a measurement instrument. As a result, there is no known degree of sensitivity of the measures or of the expected size of the measurement error.

Reviewer:  Nancy Eickelmann Review #: CR139058 (1112-1298)
Bookmark and Share
  Reviewer Selected
Featured Reviewer
Would you recommend this review?
Other reviews under "Software Architectures": Date
Software architect bootcamp (2nd ed.)
Malveau R., Mowbray T., Prentice Hall Professional Technical Reference, Upper Saddle River, NJ, 2003.  400, Type: Book (9780131412279)
Nov 11 2004
Executing SOA: a practical guide for the service-oriented architect
Bieberstein N., Laird R., Jones K., Mitra T., IBM Press, 2008.  240, Type: Book (9780132353748)
Aug 12 2008
Agile software architecture: aligning agile processes and software architectures
Ali Babar M., Brown A., Mistrik I., Morgan Kaufmann Publishers Inc., Waltham, MA, 2014.  432, Type: Book (978-0-124077-72-0)
Aug 6 2014

E-Mail This Printer-Friendly
Send Your Comments
Contact Us
Reproduction in whole or in part without permission is prohibited.   Copyright 2004™
Terms of Use
| Privacy Policy