Regular files, can currently store metadata.{height,width} (e.g. for images). We're planning to use media files to store the annotations.json WADM, and the page images for PDFs.

Some properties we need to capture are:

• access.hidden: true/false: if a file will be visible in the record landing page "System files" box
  • We don't want the pages images to be visible, since we might have 100+ of them for larger PDFs
• metadata.page: order/numbering for the page
  • This is also important for sorting in the IIIF manifest
• metadata.{type,language,encoding,charset}: one idea was to match the HTTP Accept-... headers semantics
  • These could later be used for actual Content Negotiation, filtering a list of files, etc.
• metadata.related_identifiers: in case we have multiple PDFs (or any "annotate-able") objects, we might want to link specific sets of media files to them
  • E.g. page images for paper.pdf, page images for slides.pdf in the same record
  • This could be done using PartOf or Describes relation types
• metadata.previewer: So that on a PDF file you could set e.g. "mirador" and it would default to the Mirador previewer (instead of the default PDF.js)