Feature Interations in Telecommunications Systems

Imagine commuting home from work one day, driving past the downtown stores toward the grocery store when you receive a call on your mobile phone. ItÕs your daughter, asking you to pick up milk on the way home. Now imagine that the call is from your refrigerator. This vision is not so far-fetched with todayÕs computing and communications technologies. The communications network already provides messaging services, and these can easily be extended to support such automated messaging. Global positioning systems can determine where you are. Such devices are already installed in cars to provide directions to the drivers. Determining that you are about to pass the grocery store is just a small step from this. Finally, many appliances already use chips for various functions. ItÕs not much of a stretch for appliance manufacturers to add distributed objects to enable their appliances to communicate. Perhaps the most far-fetched idea is trusting electronic appliances to manage our lives to this extent! This current issue of Computer Networks addresses the interworking of services that the above vision requires. Feature interactions are a major hindrance to proper interworking of di€erent kinds of services, such as messaging services with location detection systems and appliance services. Billing issues could be a hindrance, too. Suppose that the wireless provider, the wireline provider, and the messaging service provider are all di€erent companies, as they might well be in a competitive marketplace. Di€erent approaches to billing could make it impossible to use the services together, even if they function together exactly as they should. In 1998, the Fifth International Workshop on Feature Interactions [1] sponsored the ®rst-ever Feature Interaction Detection Contest. The goal of the contest was to compare di€erent automated tools for detecting feature interactions from the feature requirements. This special issue of COMNET contains a paper by the contest committee explaining the contest and papers from four of the contestants on their tools for detecting feature interactions. The contest represents a signi®cant advance in the understanding of what is a feature interaction and in the methods for feature interaction detection evaluating tools. It also gave us a greater appreciation of what remains to be done. In this guest editorial, we brie ̄y describe the contest and then spend some time discussing the contest results and ongoing issues.