Basic Newspaper

Use Case

Digitized Newspapers are more complicated than some other types of content as the hierarchy of Titles, Volumes and Issues are important in making them accessible. The date of publication and providing access to OCR data are also important to the viewing experience. This recipe gives an example of a basic Newspaper with two issues and links to text generated from Optical Character Recognition (OCR) software. The aim is to give a good outline of common mapping solutions for Newspapers.

Implementation notes

This recipe was put together by Nuno Freire from Europeana and was discussed by the IIIF Newspaper Community Group. The example is a cut down Newspaper with two issues containing two pages. The OCR is at page level rather than article level. The example is for a German newspaper called the ‘Berliner Tageblatt’ and is hosted by Europeana. In the IIIF domain a newspaper title is considered to be a IIIF Collection and points to each issue, represented as IIIF Manifests, in a member of the items element. An issue in this example is considered to be the unit or edition that was published on a particular date. For a more complete discussion on Newspaper hierarchies and how they map to IIIF see “Newspapers and IIIF Structures” below.

(For a full discussion of navigation by chronological order, see Navigation by Chronology.)

One important feature of newspapers is their publication date, which IIIF manifests can provide to viewers to allow navigation of issues by chronology. To achieve this we use the navDate property. This property should be added both to the items in a collection and to each issue’s manifest. Viewers are not required to make use of navDate, and viewers that do have date-aware navigation available may not default to that navigation interface.

Because manifests have no enforced truth relation to digital objects, an arbitrary temporal value can be inserted to enforce navigation order for newspaper editions. You could, for instance, use T06:00:00+00:00 in the timestamp portion of a navDate property value for a morning edition and T17:00:00+00:00 for an evening edition just to provide browse order. If you can insert the actual publication time and time zone or time zone offset for an edition, so much the better.

Linking to OCR Text

As Annotations

Digitized newspapers often have associated OCR text. To make this available inside a IIIF viewer, it needs to be in the form of one or more Annotations in the appropriate manifest, structured substantially similarly to the captions or subtitles of an A/V file. Each OCR file should correspond to a Canvas, and should be an Annotation in an Annotation Page, with

  • a motivation of supplementing,
  • the URI of the OCR file in the id property of the Annotation body, and
  • the target set to the applicable Canvas.

If an OCR file does not represent the entire Canvas, the Annotation’s target should include a fragment in #xywh Fragment Selector syntax of the portion of the Canvas the OCR represents. For more detail, see this cookbook’s entry for Transcripts, Captions, and Subtitles — General Considerations.

Linking Directly to an ALTO File

As well as linking to Annotations, it is a common use case with newspapers to link to other formats of the content, including the open XML Schema ALTO. This is achieved by using a rendering property on the Canvas, as the ALTO content is a alternative representation of the page. ALTO content differs from OCR in that it is a representation of a portion of a newspaper object, not an alternative format of the content of a portion of a newspaper. For more detail on using additional files connected to a newspaper, see the Providing Alternative Representations recipe.

Newspapers and IIIF Structures

Members of the IIIF community have arrived at the following as the best correspondences between real-world newspaper components and IIIF structures. At a minimum, a newspaper rendered as IIIF should have a title, issue, and page.

Newspaper IIIF
Summary Title / Collection Title / Curated Title / Collection (a locally defined value for user access and to address title changes per serials cataloging) Collection
Title (usually the masthead title) Collection
Volume Collection (often more useful for provenance than user interaction and presentation)
Issue Manifest (navDate recommended for presentation browsing experience)
Edition Manifest (navDate recommended for presentation browsing experience)
Section Range
Page Canvas
Page Text (OCR) Annotation List
Article Range
Illustration Range
Supplement Manifest or Range

Example

To demonstrate the files that might be included in even a basic newspaper presentation, we’ve included a number of manifests:

  1. Newspaper Title IIIF Collection - newspaper_title-collection.json
  2. Issue 1 Manifest - newspaper_issue_1-manifest.json
  3. Issue 2 Manifest - newspaper_issue_2-manifest.json

We won’t display all of these here, to keep the recipe readable. However, we’ve inclued the Manifest for the title (structured as a IIIF Collection), thinking it will provide the best initial utility. Notable lines of the Manifest are highlighted.

Viewer support for any particular feature will depend on the viewer and any customizations or extensions. Below is a table showing viewer support as of May 2024 for features noted in this recipe.

IIIF component Viewer Support (May 2024)
Collection Mirador, Annona, Universal Viewer, Clover, Glycerine Viewer, Theseus
Manifest per issue Mirador, Annona, Universal Viewer, Clover, Glycerine Viewer, Theseus
navDate Universal Viewer, Theseus
OCR in Annotations Annona, Glycerine Viewer, Theseus
ALTO via rendering Glycerine Viewer, Theseus
seeAlso for dataset version Annona, Mirador, Glycerine Viewer, Theseus

Similarly, the quality of the OCR as well as the formatting and positioning data of the Annotations containing it will determine how useful it is. IIIF Presentation v3 only provides a standard for structuring it and communicating layout to a viewer.

JSON-LD | View in Mirador | View in Annona | View in Universal Viewer | View in Clover | View in Glycerine Viewer | View in Theseus