Annotation in Digital Scholarly Editions

Annotation in digital scholarly editions (of historical documents, literary works, letters, etc.) has long been recognized as an important desideratum, but has also proven to be an elusive ideal. In so far as annotation functionality is available, it is usually developed for a single edition and cannot easily be deployed elsewhere. There is a very fundamental reason for this: the interface to the edition that the user gets to see, is only a presentation layer, usually in HTML, perhaps as a pdf or as an e-book. What the user sees is a temporary representation of the edition, the true structure of which remains 'under the hood', usually in an XML file, a set of XML files or a database. Annotation should refer to structural and stable objects (verse lines, words in lines, paragraphs, document sections, etc.) in the edition's source files, that cannot necessarily be identified from the HTML representation of the text. It is important to understand that this internal structure can be very complex, containing different sorts of editorial annotation, perhaps multiple layers of critical apparatus comparing the text to earlier or later versions, in the case of hand-written texts information about deletions in and additions to the text, and often voluminous editorial introductory material, linked to the text. An important challenge for applying a generic annotation tool for scholarly editions is thus to develop a set of agreements about how this highly complex internal structure of the edition can be represented in (or linked from) the HTML display. The annotation tool should be able to read that information and construct durable anchors for the annotations. These anchors should facilitate display of the relevant annotations in different edition display formats and should remain stable while other aspects of the edition's interface may change, depending on changes in the environment, design trends and developing technology. In our talk, we will discuss how the information about the edition's structure can be made available to annotation tools. Technologies that we will consider include RDFa, Shared Canvas/IIIF and CTS URN's. We will discuss their strengths and weaknesses in this context and propose a solution