An emerging class of technologies defined as Service-Oriented Architecture (SOA) has been heralded as the answer for inflexible IT architecture and promises to reduce operational barriers of current IT infrastructures. In SOA, loosely coupled Web services are integrated to provide dynamic digital capabilities within and across enterprise boundaries. Little research exists on development processes of information systems using Web services and against certain development metrics. One way to perform such research is to propose a development approach, identify the metrics, and embed the metrics into the technique of service composition to allow system development with desired characteristics. This paper reports an approach to information system development based on Web services composition and the metrics designed for such approaches. This approach is based on semi-automatic, interactive, and iterative Web service composition -- a hybrid technique based on developing and searching an AND/OR graph for composite services discovery while taking into consideration human judgment for solution selection and validation by interactions in an iterative way. The composition process leverages historical Web service usage data and provides helpful suggestions to the users regarding available component services. The authors propose that the metrics can investigate the characteristics of such development approaches.
Data Mining in Web Services Discovery and Monitoring