Business Rules Extraction from Large Legacy Systems

Xinyu Wang,Jianling Sun,Xiaohu Yang,Zhijun He,Srini Maddineni
2004-01-01
Abstract:Business rules are a set of conditional operationsattached to a given data result. On legacy systems, it isvery difficult to extract business rules because of theinconsistency of documentation. Some techniques havebeen presented for extracting business rules from legacysystems. But usefulness of these methods is limited whenthey are applied to large complex legacy systems.Generally, large legacy systems involve large amount ofcode, domain variables, synonym variables and businessrules, which make it more difficult to extract businessrules. This paper proposes a framework, which offersdistinct advantages over normal extraction solutions forlarge legacy systems. This framework consists of fivesteps: slicing program, identifying domain variables, dataanalysis, presenting business rules, and businessvalidation. It has been applied to a large complexfinancial legacy system which has proved to be successful.
What problem does this paper attempt to address?