DH2012: NeDiMaH workshop on ontology based annotation (pre-conference)

This session was billed as a workshop, but was really a series of presentations - sort of a mixed bag of things relating to annotation, ontologies, etc, as Rob Sanderson (who presented on Open Annotation) tweeted:

Workshop on Annotation could have done with stronger definition. From literary criticism, to POS tagging, now keyword centrality. #dh2012— Rob Sanderson (@azaroth42) July 17, 2012

I appreciated learning more about the Open Annotation Ontology from Rob Sanderson’s presentation, and was interested to hear about the merging, collaborative work between different groups working on annotation from different angles - the combination of humanities and sciences approaches seems to bode well for a robust standard that will be widely applicable. I was interested to see the complete list of the many kinds of annotations; we briefly considered using Open Annotation for the Readux/Yellowbacks prototype, and it covers everything we prototyped (bookmark, highlight, moderation) and other features we considered (comments, tagging), plus a few others we hadn’t thought of. It seems that they have really thought through a lot of potential complex cases– annotating multiple items at once; associating different types of materials (e.g., a recording of an audio performance with an image of the sheet music). I was particularly interested in the ways they are working to allow citing a specific part of a resource, e.g. referencing specific text on a page by offset and length or by quotation (the latter in an effort to allow for some textual change on the resource). The quotation case is an interesting one, because it raises a copyright issue– if people highlight or annotate the entirety of a newly published text, then the quoted text is effectively made available. This makes me think that perhaps content producers could do more to make their content more addressable; if epublishers are worried about this, then maybe they should provide better mechanisms to reference and address content.

There was a presentation about applying the OAC model for scholarly electronic editions; using OA for a variety of textual notes and FRBR to document or relate different versions of a work, but with support for collaboration, for example replies to annotations could allow for scholarly conversation about a work.

John Bradley talked about a “more informal’ interpretive model, trying to think about the scholarly process and what scholars actually want to produce at the end of their research process. He suggests that process is one of working towards an ontology (maybe), but may not get there - a cyclical process of reading, interpreting (generating concepts, adding structure to thoughts), and writing - which organizes those concepts. A personal interpretation might connect to a domain ontology, maybe - but isn’t one itself. He then demonstrated some software tools - annotating a PDF, annotation in WordHoard, and Pliny, which allows you to assemble and visually organize annotations, adding notes and comments to those notes, or making them into a concept map. As he demonstrated how Pliny worked, I tried to think about my own process and how it would or wouldn’t map to this kind of a workflow. It’s been a while since I did any real scholarship, but I do remember a sense that I tend to think differently with a pen in hand– often drafting outlines or organizing concepts with pen and paper before fleshing them out in prose. From the screenshots I’ve seen, it looks like Pliny has a somewhat dated UI, which unfortunately may keep people from using it when it might actually fit their process; I’m not convinced it would work for me, but I do think the organizing and categorizing step might work really well (for me, at least) on a large touch/surface interface. Or, better yet, integrate this kind of workflow and thinking into whatever existing software tools scholars are already using, whether that is zotero, evernote, or something else.

There were a few other presentations, of varying relevance and applicability. One presenter discussed some of the problems and limitations of this kind of ontology work (vagueness and contradictions; ontology must be agreed upon but interpretations aren’t necessarily shared; marking or documenting properties that change; modeling continuous processes or movements), but the comments and questions from other participants seemed to indicate that the community either already can handle, or is working towards being able to handle, these types of things (e.g., RDF as quads instead of triples, with context as the fourth term).