科研之友
微信
新浪微博
Facebook
分享链接
摘要
After many years of information systems development, most private and public organizations are characterized by the presence of multiple information systems, possibly distributed and heterogeneous. Heterogeneity is generally related to representation languages, support technology, and evolution strategies. The result can be a deep disintegration of data and processes spread in several information systems. Methods and tools for the analysis and comparison of the existing information systems are required, to identify replication, overlapping, bad distribution of data and processes among the existing systems, to set the basis for reengineering activities. This paper proposes computer-based techniques for the analysis of multiple information systems. Following an inherently data-oriented approach, conceptual descriptions of processes are analyzed focusing on characteristics of data manipulated and exchanged and on operations performed by processes. The proposed techniques rely on similarity criteria and metrics and on a semantic dictionary, where the knowledge on process data and operations is properly organized. Process descriptions are analyzed for the aspects related to information and operation similarity, to evaluate semantic correspondences between processes and identify activity replication and overlapping, as well as for the aspects related to interaction/cooperation, to evaluate the degree of coupling between processes and identify the type and the nature of exchanged information flows. Experimental results of applying the analysis techniques to the information systems of the Italian Public Administration are discussed.
- 出版日期1998-6
