Slaviša Marković

Refactoring OCL annotated UML class diagrams

Refactoring of UML class diagrams is an emerging research topic and heavily inspired by refactoring of program code written in object-oriented implementation languages. Current class diagram refactoring techniques concentrate on the diagrammatic part but neglect OCL constraints that might become syntactically incorrect by changing the underlying class diagram. This paper formalizes the most important refactoring rules for class diagrams and classifies them with respect to their impact on attached OCL constraints. For refactoring rules that have an impact on OCL constraints, we formalize the necessary changes of the attached constraints. Our refactoring rules are specified in a graph-grammar inspired formalism. They have been implemented as QVT transformation rules. We finally discuss for our refactoring rules the problem of syntax preservation and show, by using the KeY-system, how this can be resolved.

A graphical approach to prove the semantic preservation of UML/OCL refactoring rules

Refactoring is a powerful technique to improve the quality of software models including implementation code. The software developer applies successively so-called refactoring rules on the current software model and transforms it into a new model. Ideally, the application of a refactoring rule preserves the semantics of the model, on which it is applied. In this paper, we present a simple criterion and a proof technique for the semantic preservation of refactoring rules that are defined for UML class diagrams and OCL constraints. Our approach is based on a novel formalization of the OCL semantics in form of graph transformation rules. We illustrate our approach using the refactoring rule MoveAttribute.

Semantics of OCL specified with QVT

The Object Constraint Language (OCL) has been for many years formalized both in its syntax and semantics in the language standard. While the official definition of OCL’s syntax is already widely accepted and strictly supported by most OCL tools, there is no such agreement on OCL’s semantics, yet. In this paper, we propose an approach based on metamodeling and model transformations for formalizing the semantics of OCL. Similarly to OCL’s official semantics, our semantics formalizes the semantic domain of OCL, i.e. the possible values to which OCL expressions can evaluate, by a metamodel. Contrary to OCL’s official semantics, the evaluation of OCL expressions is formalized in our approach by model transformations written in QVT. Thanks to the chosen format, our semantics definition for OCL can be automatically transformed into a tool, which evaluates OCL expressions in a given context. Our work on the formalization of OCL’s semantics resulted also in the identification and better understanding of important semantic concepts, on which OCL relies. These insights are of great help when OCL has to be tailored as a constraint language of a given DSL. We show on an example, how the semantics of OCL has to be redefined in order to become a constraint language in a database domain.

An OCL semantics specified with QVT

Metamodeling became in the last decade a widely accepted tool to describe the (abstract) syntax of modeling languages in a concise, but yet precise way. For the description of the languages semantics, the situation is less satisfactory and formal semantics definitions are still seen as a challenge. In this paper, we propose an approach to specify the semantics of modeling languages in a graphical way. As an example, we describe the evaluation semantics of OCL by transformation rules written in the graphical formalism QVT. We believe that the graphical format of our OCL semantics has natural advantages with respect to understandability compared to existing formalizations of OCLs semantics. Our semantics can also be seen as a reference implementation of an OCL evaluator, because the transformation rules can be executed by any QVT compliant transformation engine.

Definition and correct refinement of operation specifications

Modern incremental and iterative software engineering processes advocate to build software systems by first creating a highly simplified and abstract model of the system which is then moved by applying a series of model improvements toward implementation. Models of software systems at any level of abstraction should contain, besides structural information, a precise description of the expected system behavior. This paper formalizes relations between models of the same system at different levels of abstraction, classifies approaches for describing behavior of system operations, and investigates how these system operation descriptions can be kept synchronized with frequent changes of the systems structure.