Page turner discussion

Owned by Bill Branan
Jun 16, 2017
3 min read

 

  • who's built/using a pt?
    • Natahn Talman, UCin: no pt / using IA on some instances; want to get it into
    • Rob, Stanford: 3 different pt’s; Mirador2, IIIF
    • Steven Ng, Temple: don’t have pt; experimenting with seadragon and loris
    • Jeremy M, MPub: looking for a post-DLXS publishing platform; fully encoding text
    • Princeton: handful of pt’s; kill them all slowly; IA
    • Jess Keck, Stanford: embark how to embed digital objects around sites; pull in a pt environment into an application (embed vs. destination)
    • Chris Powell: pt’s
    • Tom Cramer, Stanford: would really like Hydra to be native IIIF compatible and can swap pt’s
    • Justin Coyne, DCE: pt for Case / TEI + 1/2image; AngularJS; wrote RIIIF
    • Randal Foy, IU: home-grown pt’s; out of METS, walking TEI; what happens in Hydra
    • John Dun, IU: specialized pt part of variations musical library (score + music specific annotation tools); integration with annotation (replace variations with avalon)
    • Betsy Colman: built pt’s
    • Scott Smith, USCB: how might pt be incorporated in Hydra
    • USCB? : looking into pt
    • Neal Stewart, : used IAs
    • Peter Binkly, : legacy project with pt
    • Richard Green, Hull: 5 UK universities, share a common need for a pt; Hull has taken IA and done a proof of concept against Hydra but it’s dirty. Not really looking for a heavyweight pt, just want to work with JPEGs
    • Brian Madey, DCE:
    • Dan Brubraker-Horst, ND: IA pt
    • Evan English, BPL: rolled out a pt in hydra, blacklight component based on WDL viewer
    • Robin Ruggaber, UVA: move to IIIF compliant pt
  • What about the back-end? Special structures that you’ve needed to build into Fedora / back-end to support the pt? Want to build a hydra environment that doesn’t care about the presentation
    • Justin Coyne: TEIp5 for Case - text + images simultaneously; is that that the common use case?
      • IU has this, using XTF
      • DLXS does this
      • HathiTrust via METS
    • Richard Green: Durham have OCRed text with images; Hull envisions something more robust than its current proof-of-concept--- json data stream with item data (image sizes, etc): easy with IA if the data all maps up
    • Evan: each image is its own object in fedora; rel-ext has two relationships (has-preceeding-page and is-preceeding-page-of) -- generate an array of PIDS in sequenced objects and fed to viewer javascript
    • IIIF has two APIS --- for getting at pixels, and one JSON based about order of images to populate a gallery, pt, etc.
  • Document driving presentation vs. each page knowing its context and then being able to tell where to go prev/next
    • images are related to parent via rel-ext
    • pre-serialize the rel-ext and embedded in HTML as js
  • TEI parsed in Ruby -> compound model -> easy to map between filename and datastream node; made it into JSON this page has this image and this text
  • Are there people ooking for a lightweight reader that only works with JPEGs, vs. being driven from a separate server (IIIF)
    • RIIIF is a middle-ground because you can just mount this gem - fetches the image out of fedora, resizes, keeps the cached copy around
    • IIIF presentation API is trying to solve is the middle format that drives the browser, define that abstraction layer for plug-n-play
  • IIIF2 going to have impact? (No, not at this level of discussion.)
  • Why are we building so many pt’s?
    • lots of people are using IA
    • HathiTrust needing something more modular
    • UVA needed to wrap more rights data
  • Any grounds for us to be looking into a gem that becomes a standard way to handle this in Hydra?
    • Mirador could be a base (does tiling), extensible
      • sprint to build annotation into openseadragon
    • Harvard Project to use Mirador 2 as a pt
  • driving force for ux: 2-page book, turning pages; don’t just want sequential images; the skeumorphism is important
  • https://github.com/IIIF/mirador
  • is there a pattern to feed one of these ready made front-ends from hydra? how do you define the structure?
    • people wanting bookreaders probably have metadata with structure
  • ability to embed content in other environments, e.g. take something out of a hydra repo and embed it somewhere else where you get analytics/access/etc.