Automated mass maintenance of a software portfolio

  • Authors:
  • Niels Veerman

  • Affiliations:
  • Department of Computer Science, Vrije Universiteit Amsterdam, HV Amsterdam, The Netherlands

  • Venue:
  • Science of Computer Programming - Special issue on source code analysis and manipulation (SCAM 2005)
  • Year:
  • 2006

Quantified Score

Hi-index 0.00

Visualization

Abstract

This is an experience report on automated mass maintenance of a large Cobol software portfolio. A company in the financial services and insurance industry upgraded their database system to a new version, affecting their entire software portfolio. The database system was accessed by the portfolio of 45 systems, totalling nearly 3000 programs and covering over 4 million lines of Cobol code. We upgraded the programs to the new database version using several automatic tools, and we performed an automated analysis supporting further manual modifications by the system experts. The automatic tools were built using a combination of lexical and syntactic technology, and they were deployed in a mass update factory to allow large-scale application to the software portfolio. The updated portfolio has been accepted and taken into production by the company, serving over 600 employees with the new database version. In this paper, we discuss the automated upgrade from problem statement to project costs.