A document manipulation system based on natural semantics

This paper reports on an application of programming environments generation to structured document manipulation. We use the Centaur system as a formal tool to model and implement logical and physical document structure, logical structure editing, layout processing, format conversion, and document queries, for a sample class of documents: scientific articles containing figures. We support two particular concrete external syntaxes for logical structure: LaTEX, and Tioga (Tioga is a wysiwyg editor in the Cedar programming environment). From the formal specifications of the logical and physical structures of the article document class, and a formal specification of varied tools such as layout processing and document queries, the Centaur system automatically generates a document manipulation system including a structured editor for articles, a format converter between LaTEX and Tioga, a previewer displaying articles in their layout form, as well as management tools over a set of articles.