Dealing with structural patterns of XML documents

Evaluating collections of XML documents without paying attention to the schema they were written in may give interesting insights about the expected characteristics of a markup language, as well as bout any regularity that may span across vocabularies and languages, and that are more fundamental and frequent than plain content models. In this paper we explore the idea of structural patterns in XML vocabularies, by examining the characteristics of elements as they are used, rather than as they are defined. We introduce from the ground up a formal theory of eight plus three structural patterns for XML elements, and verify their identifiability in a number of different XML vocabularies. The satisfying results have allowed the creation of visualization and content extraction tools that are completely independent of the schema and without any previous knowledge of the semantics and organization of the XML vocabulary of the documents.