Taming the confusion of languages

  • Authors:
  • Rolf-Helge Pfeiffer;Andrzej Wasowski

  • Affiliations:
  • IT University of Copenhagen, Denmark;IT University of Copenhagen, Denmark

  • Venue:
  • ECMFA'11 Proceedings of the 7th European conference on Modelling foundations and applications
  • Year:
  • 2011

Quantified Score

Hi-index 0.00

Visualization

Abstract

Large software systems are composed of diverse artifacts. The relations between these artifacts are usually not formalized, if the artifacts use different modeling or programming languages. This hinders component-oriented development, as interfaces of exchangeable components do not capture hidden artifact dependencies. We present GenDeMoG, a tool that allows for mining inter-component dependencies beyond those explicitly specified. GenDeMoG is a generic generator-generator parameterized with a high-level system model containing dependency specifications. So, unlike the language interface mechanisms, GenDeMoG is not restricted to any given kind of links. We apply GenDeMoG to a realistic case study-an open source enterprise system, OFBiz. The experiment confirms that the stereotypical opinion about unknown dependencies across artifact types is indeed correct. Just 22 specifications allowed GenDeMoG to uncover 1737 undocumented inter-component dependencies among OFBiz components.