Open annotation notes

Owned by Bill Branan
Jun 16, 2017
3 min read

  • Champions

    • Robert Sanderson - Stanford - co-chair of W3C Annotation working group

    • Tom Johnson - interested in use of Open Annotation gem with ActiveFedora 7+ - interested in use cases

  • Round the table

    • Jon Stroop (Princeton) - works on IIIF

    • Lynette Rayle (Cornell) - writing gem using Open Annotation (not published)

    • Katherine Lynch (Temple) - have need for Open Annotation, interested in Active Triples

    • Dan Brubaker Horst

    • John Dunn (Indiana) - interests in e.g. annotation of musical scores, ethnographic videos

    • Sharon Farnel (U of Alberta)

    • Justin Simpson (Artefactual)

    • Evan English (Boston PL)

    • Hannah Frost (Stanford)

    • Robin (UVa)

    • Andrew Woods (Duraspace) - couple beta projects with Fedora 4, including Stanford - interested in best practices

    • Raymond (UVa)

    • U of Michigan

    • John Mcloan U of Michigan - Hypothesis (open access journals)

    • Nathan Tallman - archival finding aids

    • Russel Chalbet - bibliographies

  • So - 50/50: learn more / specific use cases

  • Sanderson: intro

    • Mellon grant 2009 - merged with annotation ontology 2011, led to working group

    • will be technical recommendation in two years time

    • model based on linked data: Annotation is layer on top of web, linking objects: the annotation linked to the target - all are resources (with own URIs) - so could have a video that is a comment on an image somewhere else, etc.

    • striving for flexibility: keep easy things easy while enabling the complex things

    • easy: a tag (like or +1) about some resource

    • next step: textual comment, in-line or link, in annotation document (json-ld) - can deal with parts of resources (with URI fragments or for more complex cases can specify e.g. SVG bounding polygon on an image or text bracketting (text before, target, text after)

    • extended: lists (e.g. same comment translated into different languages)

  • presume that the target will be preserved; presume that versioning can happen

  • Linked Data for Libraries - Stanford and Cornell project - for books and library materials, physical or digital - to enable discovery and reuse, building “overlay” collections

  • BIBFRAME - work going on in LD4L - contains annotations for cover art, review, etc. - LC is part of the W3C work so hope for alignment

  • Lynette: want to do virtual collections, with two kinds of tags: open comments and semantic

  • Stanford (Sanderson) - work being done by Chris Beer - using Fedora 4 to store the annotations, with gateway that accepts json-ld and uses Fedora to manage the storage of the annotations - have set aside 6 weeks for three devs to work on “Trianon” (Fedora work); and a sprint in a couple of weeks with Hypothesis people to store annotations in Trianon or Annotation Server or Harvard project - hope to be able to demo the full stack

  • Media annotation use cases (John Dunn): scholarly annotation of ethnographic video: define segments of video, attach free-form and controlled-vocab annotations - then have annotations peer-reviewed and published - current data model is very hierarchical, vs Open Annotation’s flat structure (though you could have URIs for separate sections of hierarchy)

    • annotation of musical scores, with semantics for repeated passages etc. - want to use to e.g. display passages in different colours - like bookreader annotation functionality (in epub) to allow user to highlight text in different colours without saying what they mean

  • Ideas for a Working Group

    • Tom Johnson: DPLA is writing Open Annotation into version 4 plans (to be finalized in a couple of weeks) - no definite use cases yet - interested in letting groups create on the fly groupings of objects (e.g. for exhibit)

      • do you want to pull annotations from source collection? Would cache them, assume they exist on the web.

    • Sanderson: interested in system creating and storing annotations in Hydra context - one infrastructure for all collection

    • W3C requires two interoperable implementations

    • Hannah Frost: use case to provide summary notes in place of full transcription of a text

    • Nathan Tallman: user annotations on finding aids - “this is what’s really in this folder”

    • Robin: same for images

    • (Controlled vocab? or just free text? Tom: want to do controlled vocab early; Rob: controlled vocab is phase 2)

    • Actual commitment of programmer hours before next Hydra Connect

      • Tom

      • Lynette

      • John Dunn

      • Kate

    • Action item: Rob to pull people together