Automating XML document Transformations: A conceptual modelling based approach

The growing use of XML mark-up language has made a large amount of heterogeneous XML documents widely available. As the number of applications that utilize heterogeneous XML documents grows, the importance of XML documents transformations increases greatly. A serious obstacle for translating directly between two XML documents, using languages like XSLT, is that a mapping between the two XML representations needs to be carefully specified by a human expert. Current research attempts to address this problem by proposing algorithms to automate aspects of XML schemas matching task. In this paper, we identify two major problems encountered when current matching algorithms are used in the context of XML documents transformations. The first problem concerns possible scalability problem due to the diversity of schema constructs. The second problem deals with the need to perform semantic matching. We argue in favour of conceptual modelling as solution to avoid such problems. We introduce a new layered model for XML schemas, called Layered Interoperability Model for XML Schemas (LIMXS). LIMXS offers a semantic view for XML schemas through the specification of concepts and semantic relationships among them. We will show how our model transforms the matching algorithm into a dynamic and incremental algorithm that provide semantic mappings and helps to automate the transformation process.