Verification and validation in the context of domain-specific modelling

Abstract

The utilisation of Domain-Specific Modelling (DSM) in software development has a significant positive impact on productivity. The productivity increase is caused by the utilisation of modelling languages and generators that are especially suitable for a specific problem domain instead of those designed for solution domains. The prerequisite for this significant productivity increase is that the languages and the automation function correctly. To ensure the suitability of the languages and tools, we need to be able to use the verification and validation (V&V) techniques in the context of DSM. In this position paper we study what V&V actually stands for in this particular context and what the current means are for performing V&V. We found that although there are some means available for verification, comprehensive methods still do not exist. For validation, we believe that maintaining a bidirectional trace link between requirements, models and the generated deliverables is a promising approach to significantly facilitate the validation process.

Continuous assessment of software traceability

“…For the first time, a comprehensive traceability assessment method is presented. Recently, Merilinna and Pärssinen lamented the fact that no comprehensive traceability assessment method exists [Merilinna and Pärssinen 2010], which emphasizes the relevance of this contribution.…”

“…This traceability gap demands for adequate traceability assessments to detect these gaps [Merilinna and Pärssinen 2010]. Researchers argue that traceability must be purposed and trusted ].…”

Natural-language Scenario Descriptions for Testing Core Language Models of Domain-Specific Languages

“…Such a procedure requires a generic, proxy metamodel from which the test models are instantiated. Metamodel validation approaches (3) employ model-constraint expressions (e.g., specified via the OCL) to express test cases on metamodels (e.g., specified as invariants), defined at the level of the corresponding metametamodel (Merilinna and Pärssinen, 2010). Our approach primarily extends metamodel validation techniques to provide an explicit scenario abstraction, both at the requirements and the testing level.…”

“…The same holds true for metamodel-test models (i.e., simulating a set of valid instance model alternatives), which require the full domain model under test to be specified (Sadilek and Weißleder, 2008;Cicchetti et al, 2011). Metamodel validation approaches (e.g., model-constraint evaluations) employ formal expression languages (e.g., OCL), but do not consider the structure of non-executable requirements specifications (Merilinna and Pärssinen, 2010). Moreover, approaches exist for tracing requirements (Winkler and Pilgrim, 2010) and testing natural-language statements (Gervasi and Nuseibeh, 2002;Yue et al, 2013), but lack an integrated model-driven tool-chain to combine both, domain modeling capabilities and requirements specification/validation.…”

Implementing modular domain specific languages and analyses

“…Merilinna and his colleagues [6] state that verification in the context of domain specific modeling comprises the following aspects: verification of the meta-model which captures the domain semantics, verification of the generators which maintain the semantic equivalence between input and output models and verification of textual generators. Our approach for building modular analyses based on modular sublanguages lowers the verification efforts since it can be done only once at a language module level even if that language module is further embedded in different DSLs.…”