The Open Annotation Collaboration Model for Linked Data

Canberra.
The next speaker at DHA 2012 is Anna Gerber, whose interest is in open annotation for electronic editions. She defines annotations as additional information attached to a digital resource or part of the resource, which do not modify the original content of the resource itself.

Such annotations traditionally exist as footnotes, endnotes, glossary entries, or in other forms, providing descriptions, explanations, or justifications for particular textual or formatting choices (especially in critical editions); they may also be used to link in secondary material. But such annotations – especially in digital editions – can also be used for comments, questions, and replies; that is, to sustain a dialogue around the primary text.

There are a number of approaches to digital annotation, including TEI annotation schemes, offline annotation (in PDF, XML, eReaders), Web-based annotation (e.g. through tagging or commenting), and community-driven modelling systems (Annotea, the Annotation Ontology, the Open Annotation Collaboration, and others).

Anna has been involved in the Open Annotation Collaboration, which aims to establish some level of interoperability between online annotation systems. This is based on RDF and linked open data principles. In RDF, the key concept is a triple (subject, predicate, object), which lends itself to establishing graph-based data structures. RDF schemata provide defined vocabularies and semantic models.

Linked data use URIs as identifiers, or indeed, de-referenceable HTTP URIs; they use semantic Web standards, and they include links to related resources in order to create a Web of linked data.

The OAC model, then, translates this into a system that includes annotations, bodies, and targets; additional properties (dates, authors, etc.) can also be attached to annotations, and annotations can be about specific segments of a given resource (words, sentences, images), rather than just for the overall resource itself. In total, this is a system for attaching structured metadata to the original resource.

Interoperability challenges in all this include the migration and querying of annotations across versions; the targetting of specific resource segments regardless of document format (PDF, HTML, …); the targetting of variants; the problem of digital surrogates for facsimile versions of the original resource; the identification of abstract entities (shared vocabularies, identification guidelines, etc.); and functional interoperability. A technical infrastructure for this, lorestore, has been released via Github, and a new project, AustESE, will further this work for the authoring of electronic scholarly editions.