Manifest Reconstruction

Jeffrey C. Witt (Loyola University Maryland) | @jeffreycwitt

March 16, 2017

Slide Deck:

### The E-codices Manifest []( or []( View this in a mirador viewer []( and you will see images coming from Switzerland and a table of contents, search service, and transcriptions (at least on folio 107v) coming from Baltimore.
### Minting of Canvas Ids ``` { "@id": "", "@type": "sc:Canvas", "label": "bindingEa", "height": 6496, "width": 4872, "images": [{ "@id": "", "@type": "oa:Annotation", "motivation": "sc:painting", "on": "", "resource": { "@id":"", "@type": "dctypes:Image", "format": "image/jpeg", "height": 6496, "width": 4872, "service": { "@context": "", "@id": "", "profile": "" } } }] } ```
### Identifying canvases IDs is not as daunting as it might seem. It usually involves identifying the canvas base `` and then recognizing the changing pattern for each canvas `kba-WettF0015_e005a.json`. Note - Bodleian manuscripts are using hashes, so there will be no predictable pattern for associating canvas ids with folio pages. The point - The publication of supplemental material requires the collection of canvas ids and their association with the published supplemental data.
### Manifests (for the time being) need to publish an endpoint where they can receive updates about new information ```json service": [{ "@context": "", "@id": "", "profile": "", "label": "e-codices Webmention Service" } ``` In the future, there might be a central hub or central inbox that would negate this requirement. For the purposes of a present demo, this step can be skipped if the listening endpoint is already known.
Once the announcement has been made, an institution can incorporate this data by dynamically constructing a manifest from known distributed sources. In the following manifest: []( See lines 61-66 of ecodices manifest from search block. Compare to: []( See lines 5518-5545 for Transcription Inclusion Compare to: []( See lines 9760 and following for Table of Contents of Range List Compare to: [](