Unification of XML Documents with Concurrent Markup

Abstract

An approach to the unification of XML (Extensible Markup Language) documents with identical textual content and concurrent markup in the framework of XML-based multi-layer annotation is introduced. A Prolog program allows the possible relationships between element instances on two annotation layers that share PCDATA to be explored and also the computing of a target node hierarchy for a well-formed, merged XML document. Special attention is paid to identity conflicts between element instances, for which a default solution that takes into account metarelations that hold between element types on the different annotation layers is provided. In addition, rules can be specified by a user to prescribe how identity conflicts should be solved for certain element types.

Integrated Linguistic Annotation Models and Their Application in the Domain of Antecedent Detection

“…The different annotation layers have been combined using markup unification which allows the combination of two XML annotation layers into a new XML instance [55]. The analysis of the research questions is based on the unified annotation data.…”

“…a directed acyclic graph structure (GODDAG, [45]) over the XML inherent tree) and its notation. In addition, the Prolog fact base approach discussed by [54,55] or adding delay nodes to the XQuery 1.0 and XPath 2.0 Data Model (XDM) as virtual representation of nodes proposed by [29] allowing different nodes to share children describe other non-XML based specifications. Furthermore, XMLbased specifications that follow the Annotation Graph paradigm [4], such as NITE [6], the Potsdamer Austauschformat für Linguistische Annotationen (PAULA, cf.…”

“…Although it is possible to use the character stream of the primary data as coordinates to align different annotation layers (cf. [55]), often changes to the primary data are introduced during the annotation process (in terms of added or deleted whitespace) raising further issues.…”

“…It is an XML-based successor of the Prolog fact base format for the storage and analysis of multiple annotated texts described in [55] and [19]. It follows a standoff annotation approach but combines all annotation levels that belong to the same primary data in a single XML instance, following the formal model of a multi-rooted tree.…”

SusTEInability of linguistic resources through feature structures

“…Concerning the Operation of unification, the result of the unification of two fea ture structures can be intuitively conceived of as the fusion of the information contained in both feature structures, if the respective information packages are compatible with each other. Witt et al (2005) describe an application of unification for XML documents with concurrent markup.…”