Projects of the Experimental Software Engineering Group at the University of Maryland

Experience Domain Analysis

Approach

Preliminary versions of three methods for characterizing domains have been developed:
  • a historical data-based approach using Optimized Set Reduction (OSR) to identify domains, or groups of projects which share relevant characteristics. OSR helps to determine what project characteristics are relevant based on data on the "success" of using the reuse candidate on that project. These characteristics can then be used to assess a new project and whether it is likely to be able to successfully use the reuse candidate.
  • a qualitative method which uses subjective ratings from experts to identify the relevant characteristics of a domain relative to the experience being reused. Experts who have had experience using the reuse candidate successfully on previous projects are interviewed and they are asked both which characteristics were relevant and how relevant they were. These ratings are combined to produce guidance which can be applied to a new project to determine if it can successfully reuse the reuse candidate.
  • an approach using fuzzy logic to characterize a domain with the use of fuzzy inferences. Expert opinion is gathered, as with the second approach, but it is modeled as fuzzy propositions describing relationships between project characteristics and successful use of the reuse candidate. Also a fuzzy "reusability" function is formulated, which can be evaluated using characteristics of a new project to determine whether the reuse candidate is likely to be used successfully.

We have the opportunity to try at least one of these approaches in a case study currently underway.

Project Status

Active - Case study underway - transferring the GQM methodology from the Software Engineering Laboratory (SEL) at Goddard Space Flight Center, to be reused by the Rapid Development Laboratory at Johnson Space Center.

Results

No results at this time.

Top Level Page


<-Back to ESEG Home Page


Last updated: March 20, 1998 by Carolyn Seaman