Executable model composition: a multilevel approach

  • Authors:
  • Carlos Rodríguez;Mario Sánchez;Jorge Villalobos

  • Affiliations:
  • Universidad de los Andes, Bogotá, Colombia;Universidad de los Andes, Bogotá, Colombia;Universidad de los Andes, Bogotá, Colombia

  • Venue:
  • Proceedings of the 2011 ACM Symposium on Applied Computing
  • Year:
  • 2011

Quantified Score

Hi-index 0.00

Visualization

Abstract

Modularizing concerns is a common strategy to lower application complexity: it results in modules that are easy to maintain, to adapt, and to replace. In some cases, these modules can also be expressed with very expressive concern-specific languages that manage high level concepts. On the other hand, modularization also requires tools and languages to describe the relations and interactions between concerns and to reconstruct the full application semantics. Ideally, these descriptions should be written using languages that work with the same concepts than the concern specific languages. Cumbia is a metamodel-based platform to build applications that support multiple concern specific languages. However, the composition and coordination mechanism between concerns depends on a too low level language called CCL: since it is built around very basic and generic coordination primitives, using the language requires a strong technical knowledge of the platform, and it is difficult to implement tools that offer meaningful validations on compositions. In this paper we propose a composition metamodel to characterize, at metamodel level, the composition and coordination of Cumbia models. By adopting this solution, the mentioned problems are solved, and domain experts can have a more active role in the creation of models for Cumbia-based applications.