Abstract
Users of online digital editions have a need for annotation support to contribute explanatory material that complements what is already available in the edition itself, for purposes of private study or for publication in conjunction with a scholarly article (Boot, 2009, Robinson, 2004, Siemens et al., 2012). One challenge for browser-based annotation support is anchoring the annotation to a specific location in the digital edition, as the browser typically only has an HTML representation of the edition that describes the page layout.
Recently we proposed an ontology-based approach to describe digital editions and their components to a browser-based annotation tool so that it understands and can reason about what it is annotating (Boot et al. 2017, Boot and Koolen 2018). While the ontology was formulated in order to support principled and robust targeting of annotations, it should have a much wider applicability. In this paper, assuming a TEI context, we look at the integration between this ontology and TEI/XML files underlying the edition. This implies the need for TEI files to refer to the annotation ontology: the fragments of the edited text have to be assigned URIs, they have to be assigned a class, and their mutual relationships have to be defined in terms of the properties described in the ontology. The edition server will have to provide this information, as RDF, to the client application.
The question that we discuss in this paper is: on the basis of which information is the edition application going to make these assignments? How does it know to which classes certain text fragments belong, which URIs to assign and how to name the property that connects, say, a chapter and its paragraphs? If the user interface of the edition is generated from a TEI source(s), the natural answer to that question would be to store the relevant information in, or at least with, these TEI sources. Therefore, we need to overlay the graph model describing the edition and its content on the (hierarchical) TEI XML. There are different ways of doing that, and the main section of this paper is a practical discussion of the pros and cons of the different approaches.
The aim of this paper is not to give a final answer to the question of how and where to store the source information for the linked data necessary to facilitate interoperable annotation on a digital edition; rather we discuss a number of options, with their advantages and disadvantages.
Recently we proposed an ontology-based approach to describe digital editions and their components to a browser-based annotation tool so that it understands and can reason about what it is annotating (Boot et al. 2017, Boot and Koolen 2018). While the ontology was formulated in order to support principled and robust targeting of annotations, it should have a much wider applicability. In this paper, assuming a TEI context, we look at the integration between this ontology and TEI/XML files underlying the edition. This implies the need for TEI files to refer to the annotation ontology: the fragments of the edited text have to be assigned URIs, they have to be assigned a class, and their mutual relationships have to be defined in terms of the properties described in the ontology. The edition server will have to provide this information, as RDF, to the client application.
The question that we discuss in this paper is: on the basis of which information is the edition application going to make these assignments? How does it know to which classes certain text fragments belong, which URIs to assign and how to name the property that connects, say, a chapter and its paragraphs? If the user interface of the edition is generated from a TEI source(s), the natural answer to that question would be to store the relevant information in, or at least with, these TEI sources. Therefore, we need to overlay the graph model describing the edition and its content on the (hierarchical) TEI XML. There are different ways of doing that, and the main section of this paper is a practical discussion of the pros and cons of the different approaches.
The aim of this paper is not to give a final answer to the question of how and where to store the source information for the linked data necessary to facilitate interoperable annotation on a digital edition; rather we discuss a number of options, with their advantages and disadvantages.
| Original language | English |
|---|---|
| Publication status | Published - 2019 |