On a chain of transformations for generating alloy from NL constraints
Imran Sarwar Bajwa, Behzad Bordbar, Kyriakos Anastasakis, Mark Lee · 2012
Multi-Paradigm Modelling uses models from multiple domains to leverage the tools, techniques and expertise provided by each of the individual domains. Recent advances in model transformation technology allow automated production of one model from another to improve the application of multi-paradigm techniques. Systems development starts with the requirements gathering phase, which usually comprises of a textual description of the system requirements provided in Natural Language (NL). It is therefore evident that there is a clear scope for incorporating NL Processing techniques in Multi-Paradigm Modeling. However, using NLP methods pushes the boundaries of Multi-Paradigm Modeling to an extreme; indeed NLs are inherently ambiguous and open to interpretation. In this paper, we propose a novel approach based on standards (such as SBVR) that can cope with syntactic and semantic ambiguities in NL specifications and can map them to formal languages such as Alloy. The tool implementing our approach is currently the only available tool for translating NL specifications to formal languages such as Alloy, etc.