Hydra Newspapers Interest Group Call: 2017-02-02

Owned by Bill Branan
Jun 16, 2017
2 min read

Time: 1:00 PM EST / 10:00 AM PST

Call-In Info: 712-775-7035 (Access Code: 960009)

Moderator:  Eben English (Boston Public Library)

Notetaker: TBD (Etherpad link: https://etherpad.wikimedia.org/p/Hydra_Newspapers_Interest_Group_Call__2017-02-02)

Attendees:

Agenda:

  1. BPL-Utah grant update
    1. submitted on january 13, thanks for looking at the draft and offering comments
    2. should hear something April-ish

  2. PCDM discussion continued
    1. Examples:
      1. U. of Maryland: https://wiki.duraspace.org/download/attachments/77447979/maryland2016hyrdraconnect.pdf
      2. Nat'l Library of Wales:  PCDM Mapping for Welsh Newspapers (NLW)
    2. memberOf vs. relatedObjectOf
      1. Discussion inclusive but heading towards memberOf for it sounds like
        1. POST-MEETING UPDATE 03 Feb 2017: U. Maryland has decided to go with what seems like the consensus view (pcdm:hasMember) of the best fit for relating articles to issues.
      2. Use Article object class to distinguish from Page
      3. UM: bibo:Article (http://purl.org/ontology/bibo/Article)
      4. could use pcdm:Range?
    3. connecting Article objects to Pages
      1. NLW: mapping shows article object but not really a relationship to the page object, just to the issue
        1. IANA describedBy
      2. For locating specific articles, Maryland is planning to use search of OCR inside the issue as a secondary lookup
        1. Source data from vendor has article list, but does not include relationship of articles to image files
    4. use of Works extension models
    5. Not lots of replies from the pcdm group but no plans to get rid of works extension or range class
    6. pcdm:Range
      1. pcdm:Range is a section of the work corresponding to IIIF range
        1.   pcdm:Range is a sub-class of object
        2. pcdm:Range can have member filesets, seems like we'd have member page object that then have member filesets? Discuss more.
      2. Page object relationship to article: would need to be ordered
        1. Would proxies be necessary here? May not be use case for alternate ordering of small sequence.
      3. IIIF Range/Sequence/Canvas
        1. http://iiif.io/api/presentation/2.0/#ranges
        2. Don't know if there's anything indicating canvas order
        3. Sequence of canvases but the canvases don't say anything about order
        4. Most people are generating IIIF dynamically. Maybe storing it  you can have more relationships and modeling? Unsure.
        5. Shared canvas ontology does have sequence class
          1. http://iiif.io/model/shared-canvas/1.0/
    7. Eben will go back with more questions to the PCDM google group re: Article-Page relationship

  3. List of requirements: Features & Requirements
    1. keep adding feature requests and indicating interest
    2. Eben will send to hydra lists to get more involvement

  4. Intel sharing from other groups
    1. IIIF Newspapers Interest Group
      1. starting a Text Granularity Technical Group to deal with searching for text within images
        1. https://docs.google.com/document/d/1wTxgcj-AlAE3KwcxP59mTZhOOQKkDEaqwVK_NHOIRvc/edit?usp=sharing
        2. talk about what type of text search results to return - paragraph, line or word level
        3. elminate need for API to ALTO data
    2. Hydra plugins group
      1. https://github.com/projecthydra-labs/hydra_plugins_wg
      2. current discussions in that repo

  5. Next meeting: Thursday March 2, 1 PM EST