An Extensive Catalog of Operators for the Coupled Evolution of Metamodels and Models

Abstract

Abstract. Modeling languages and thus their metamodels are subject to change. When a metamodel is evolved, existing models may no longer conform to it. Manual migration of these models in response to metamodel evolution is tedious and error-prone. To significantly automate model migration, operator-based approaches provide reusable coupled operators that encapsulate both metamodel evolution and model migration. The success of an operator-based approach highly depends on the library of reusable coupled operators it provides. In this paper, we thus present an extensive catalog of coupled operators that is based both on a literature survey as well as real-life case studies. The catalog is organized according to a number of criteria to ease assessing the impact on models as well as selecting the right operator for a metamodel change at hand.

Metamodel and Constraints Co-evolution: A Semi Automatic Maintenance of OCL Constraints

“…For example, move property is a complex change where a property is moved from one class to another via a reference. This is composed of two atomic changes: delete property and add property [9].…”

“…Those elements represent the core features of a metamodel in the EMF/Ecore [24] and the MOF [20] standards. In the literature, over sixty complex changes are proposed [9]. Among them, we focus on seven complex changes: move property, pull property, push property, extract super class, atten hierarchy, extract class, and inline class.…”

“…It depicts an evolution of a metamodel where the property depth is deleted from the superclass Component and added to the subclass Composite, which ts the denition of a "push property" [9]. The rst two constraints become invalid because depth is no more accessible in Component.…”

Hybrid Approach for Metamodel and Model Co-evolution

“…atomic). In fact, composite or compound changes have been already considered in previous works like [12,13]. But, we envision tackle the problem differently.…”

Negotiated Grammar Evolution.

“…Many authors -in particular, Herrmannsdörfer et al [HBJ09,HVW11], Cicchetti et al [CREP08], Wachsmuth [Wac07] -have previously considered, proposed or analysed metamodel transformation operators that are somewhat similar to grammar transformation operators. In this paper, we have limited ourselves to grammar evolution not only because grammars are considered somewhat simpler than arbitrary metamodels, but also because metamodel evolution scripts are traditionally written in a more adaptable way, so they suffer less from the problem we are solving here.…”