The research of Dr. Eberlein is part of the Software Lab in the Department of Electrical and Computer Engineering and the  Software Engineering Research Network of the University of Calgary.  He is also a member of the Alberta Software Engineering Research Consortium ASERC.

Dr. Eberlein's research area is requirements engineering, which has traditionally been seen as the part of software engineering that focuses on the specification of the system to be developed. Unfortunately, RE has proven to be much more challenging than initially anticipated. While huge progress has been made in design and implementation technologies over the last years, RE is still one of the key hurdles when it comes to developing complex systems.

So far, his research has focused on the following areas:

1)      Process Support for Requirements Engineering:

In the past, his work focused on the application of artificial intelligence (AI) techniques to the requirements engineering process. He developed an expert system called RATS that actively supports the requirements engineer during the specification of telecommunication service development.

This research is currently being expanded into two directions:

(a) We are currently in the process of generalizing our previous work in order to have a more universal tool that helps with RE process development. The methodology called MRETS has already been applied to several industrial case studies. Direct project comparisons show that RE process development using MRETS has resulted in significantly higher project success.

(b) Further investigation into the use of AI to RE. The fuzzy nature of RE makes the application of AI very difficult. At the moment it is not yet certain how long it will take until industrially usable results can be found.

2)      Distributed requirements negotiations:

During this research (carried out with Dana Damian), a methodology was developed that provides guidelines for the design of empirical studies of requirements negotiation situations and of computer support in distributed development teams. The methodology was applied to the design of an empirical study in a controlled environment. Contrary to traditional wisdom, groups in face-to-face meetings performed no better than those in computer-supported settings. The empirical evidence also indicates that (1) the distributed communication of stakeholders contributes to a more rational approach to conflict resolution, and (2) that the human facilitation of distributed settings is possible. This research has resulted in a best paper award on the 4th IEEE International Conference on Requirements Engineering (ICRE 2000).

3)      Feature Interaction:

Component-based development has enjoyed increasing popularity. One of its advantages is that it allows the reuse of already developed and proven components thus speeding up the development. However, interactions between components (or features) have caused serious problems. In the past, such interactions required the employment of experienced developers who can point out possible feature interactions before the system is tested. This approach has proven to be very unreliable. Another possibility is to use formal models to see if the components interact. However, creating these models requires significant resources. We have developed a semi-formal approach to feature interaction detection called IRIS and have applied it to several case studies from the control domain, smart homes domain and telecommunications domain. So far, the results are very promising. In future, we will work on the integration of our semi-formal approach with formal and informal methodologies to feature interaction detection.

This page was last edited on March 04, 2007.