FST Graduate Research Topic

Research Topic: Domain Analysis in Software Product Line Feature Models
Suggested By: Ebrahim Bagheri
Last Update: February 5th, 2010

Typically, feature models represent variability and commonality in a software product family. The general understanding is that feature models are a result of a thorough domain analysis process in a given domain and hence are a comprehensive representative of the possible features of a product family. As the assumption goes, comprehensiveness of a feature model is fundamental for the development of suitable instances of the product family. Therefore, incomplete feature models that do not fully cover the problem domain, or fail to reflect the information gathered from the domain analysis will significantly deteriorate the quality of the products of the domain. For this reason, in this research work, we will study how feature models can be incorporated into and hence be developed simultaneously as a result of the domain analysis process in order to maintain desired quality standards for feature models such as comprehensiveness, correctness, etc. A possible approach that we will be investigating is to couple the results of goal-oriented analysis frameworks such as i* or GRAIL/KAOS with feature modeling technology. This will allow us to inter-relate domain goals and their operational scenarios with domain features, which will facilitate a thorough coverage analysis over the features available in a feature model. This project will be done in the AU semantic technologies lab under the supervision of both Dr. Ebrahim Bagheri and Dr. Dragan Gasevic as well as in collaboration with an energetic team of postdoctoral fellows and PhD students associated with this lab