Complete Specification Coverage in Automatically Generated Conformance Test Cases for TGG Implementations (bibtex)
by , ,
Abstract:
Model transformations can be specified using an operational or a relational approach. For a relational approach, an operationalization must be derived from the transformation specification using approved formal concepts, so that the operationalization conforms to the specification. A conforming operationalization transforms a source model S to a target model T, which is moreover related to S according to the relational transformation specification. The conformance of an operationalization with its relational specification must be tested since it is not certain that the formal concepts have been correctly realized by the implementation. Moreover, transformation implementations often perform optimizations, which may violate conformance. The Triple Graph Grammar (TGG) approach is an important representative of relational model transformations. This paper presents an extension of an existing automatic conformance testing framework for TGG implementations. This testing framework exploits the grammar character of TGGs to automatically generate test input models together with their expected result so that a complete oracle is obtained. The extension uses dependencies implicitly present in a TGG to generate minimal test cases covering all rules and dependencies in the TGG specification if the TGG is well-formed. In comparison to the previous random approach, this guided approach allows more efficient generation of higher quality test cases and, therefore, more thorough conformance testing of TGG implementations. The approach is evaluated using several TGGs, including one stemming from an industrial case study.
Reference:
Complete Specification Coverage in Automatically Generated Conformance Test Cases for TGG Implementations (Stephan Hildebrandt, Leen Lambers, Holger Giese), Chapter in Theory and Practice of Model Transformations (Keith Duddy, Gerti Kappel, eds.), Springer Berlin Heidelberg, volume 7909, 2013.
Bibtex Entry:
@InCollection{HLG2013,
AUTHOR = {Hildebrandt, Stephan and Lambers, Leen and Giese, Holger},
TITLE = {{Complete Specification Coverage in Automatically Generated Conformance Test Cases for TGG Implementations}},
YEAR = {2013},
BOOKTITLE = {Theory and Practice of Model Transformations},
VOLUME = {7909},
PAGES = {174-188},
EDITOR = {Duddy, Keith and Kappel, Gerti},
SERIES = {Lecture Notes in Computer Science},
PUBLISHER = {Springer Berlin Heidelberg},
URL = {http://dx.doi.org/10.1007/978-3-642-38883-5_16},
ABSTRACT = {Model transformations can be specified using an operational or a relational approach. For a relational approach, an operationalization must be derived from the transformation specification using approved formal concepts, so that the operationalization conforms to the specification. A conforming operationalization transforms a source model S to a target model T, which is moreover related to S according to the relational transformation specification. The conformance of an operationalization with its relational specification must be tested since it is not certain that the formal concepts have been correctly realized by the implementation. Moreover, transformation implementations often perform optimizations, which may violate conformance.

The Triple Graph Grammar (TGG) approach is an important representative of relational model transformations. This paper presents an extension of an existing automatic conformance testing framework for TGG implementations. This testing framework exploits the grammar character of TGGs to automatically generate test input models together with their expected result so that a complete oracle is obtained. The extension uses dependencies implicitly present in a TGG to generate minimal test cases covering all rules and dependencies in the TGG specification if the TGG is well-formed. In comparison to the previous random approach, this guided approach allows more efficient generation of higher quality test cases and, therefore, more thorough conformance testing of TGG implementations. The approach is evaluated using several TGGs, including one stemming from an industrial case study.
}
}
Powered by bibtexbrowser