Processing Lax XML Element Trees

Abstract

When processing HTML tables within XML documents, we generally find the lax element schema for the HTML table specification increases the complexity of the processing system. By first transforming HTML tables to a normalized representation we avoid the need for different logic paths deep inside our code. Ensuring there is a single normalized element hierarchy for any given table also simplifies the task of comparing different versions of each table. This paper describes the design and implementation of XSLT that transforms lax HTML table element structures within an XML document to conform to a strict content model.