Abstract: This paper presents a novel platform based methodology for the transformation of Document Type Definition (DTD) models using an attributed grammar. The proposed approach consists of two main phases. In the first phase, we define a new representation of the DTD called $$DTD_{xml}$$, which is created using the attributed grammar theory from the compiler design. In the second phase, this intermediate model $$DTD_{xml}$$ is used to facilitate the transformation of the DTD into arbitrary target models via XSLT stylesheets. To enable this transformation, we introduce an algebraic grammar for the DTD language that can syntactically recognize and parse any DTD specification. This grammar is then used to define an attribute system that can be used to semantically annotate DTD documents and transform them into an intermediate XML based representation. This structured representation, $$DTD_{xml}$$, serves as input for XSLT transformations that generate the corresponding target relational schemas or other model formats. The main contribution of this work is the decomposition of the model transformation process into two distinct phases: a structuring phase, in which the source DTD is rewritten into an XML syntax using attributed grammar techniques, and a target model generation phase, in which the normalized intermediate representation is translated into the desired model using Extensible Stylesheet Language Transformations (XSLT). This modular separation improves the reusability, scalability and extensibility of model transformation processes. It allows the compilation process to be carried out once, regardless of the number of transformations required. Each transformation is then defined via an independent stylesheet, so that a new lexical and syntactic analysis is no longer necessary.

Keywords: Attributed Grammar, Model Conversion, DTD, Relational Schema, Theory of Computation, XSLT

Pages: 157-180