Language-based document processing

This paper proposes an application of programming environments generation to structured documents manipulation. We use Centaur as a formal tool to model and implement logical and physical structure, logical editing and layout processing, document analysis, re-use and conversion for a sample class of documents : scientific articles including equations and figures. To make connections with real document systems, we choose to give two particular externals forms to the logical structure : Tioga source and Latex source. From the specifications of the logical and physical structures of the Article document class on one hand, and, on the other hand, the specification of the layout processing (viewed as its semantics according to the Tioga or the Latex layout model) and other semantic tools, the Centaur system automatically generates structured environments for Tioga and Latex documents and conversions between them.