Abstract :
COBOL systems, which often rely on legacy sequential file systems, still form a critical part in core business systems, particularly transactional systems. Understanding these systems, through domain analysis, is a critical part of any program code/data reengineering effort. Our approach involves first translating COBOL into a Wide Spectrum Language (WSL) and then subjecting this translation to a domain analysis in order to determine its proper data structures, semantic data meaning, and data constraints. The domain analysis also helps us understand the programpsilas purpose, its data structures, and some of its business rules. Once the domain analysis phase finishes, the legacy records can be translated into a relational databasepsilas multiple, linked SQL tables with proper data constraints. After conversion to a relational database, the program code is parsed in order to identify and replace file read/write statements with their equivalent SQL statements. With this replacement, redundant code associated with the master-file update file approach and temporary records can be eliminated and, thus, code size is greatly decreased.
Keywords :
COBOL; data structures; software maintenance; systems re-engineering; COBOL system; Wide Spectrum Language; business systems; data constraint; data reengineering; data structure; domain analysis; legacy sequential file systems; linked SQL table; program code; reengineering process; relational database; semantic data meaning; transactional systems; Application software; Business process re-engineering; Computer applications; Costs; Data structures; File systems; Logic; Relational databases; Software maintenance; Transaction databases; Wide Spectrum Language (WSL); data reengineering; domain analysis; legacy system;