Heritage objects in place
Several Norwegian collections publish their objects through open interfaces: the National Library through its catalogue and IIIF manifests, and most museums, the National Museum among them, through DigitaltMuseum. This page explores how it is possible to integrate such data and media on an external website.
A photograph at the National Library
An object at a city museum, through DigitaltMuseum
A painting at the National Museum, through DigitaltMuseum
The same fiddle at Europeana, and the term at KulturNav
The tunes written down, at the National Library
The National Library’s music manuscripts are the one part of its collection that this page can show without asking anyone. 7,678 of the 11,150 are freely viewable, and every one sampled carries CC BY-NC-ND 4.0 and a IIIF image service, the same kind of service as the photograph at the top of this page. Among them are the folk music collections gathered from the 1840s on, 2,624 manuscripts from the six collectors the library lists: Lindeman, Crøger, Sande, Elling, Sandvik and Groven.
This one is a set of tunes for the same instrument the objects above stand for. It runs to 1,373 pages, so the viewer below turns them, and only the page you ask for is fetched, which is what makes a manuscript of this size cheap to put on a page at all.
Turning a page costs one request for the page’s description and then its tiles. The collection’s image server sends no caching instructions at all, no cache-control, no etag and no last-modified, so a browser has nothing to decide freshness from and asks again every time. This page therefore keeps each description once it has it: going back to a page you have already seen now costs no description request, where before it cost one. The tiles are the collection’s to control, and a cache-control header on their side would save them the traffic as much as it would save the reader the wait.
A composer and performer, in the shared authority register
The National Library and Sikt keep a shared register of people and organisations, Felles autoritetsregister, which libraries use to say which person a book or a manuscript is by. Its open interface gives, for a person, the established form of the name and the dates, the other forms the name is written in, a short note on who the person is, and identifiers for the same person elsewhere: a persistent handle, ISNI, VIAF, and others such as Discogs. It holds no biography and no pictures.
The National Library’s catalogue does not yet carry the register’s identifier on each item. The include therefore finds the person’s works by the established name and dates together, which is how the catalogue records a person, and links each one by its URN where it has one. Wikidata stores the register’s identifier as P1015, so the person’s Wikidata item is found the same way as for the objects above.
Here is the violinist and composer Ole Bull, with one of his manuscripts below.
An audio file, from a Norwegian museum
Norwegian museums publish sound through DigitaltMuseum, and 1,577 records carry a recording. Of those, 592 have a Creative Commons licence. The file is served as an ordinary address, so the same include that shows a picture can show a recording. Here is a telling of how the fiddle tune Refshaugen got its name, from a record that names its reader and its licence.
An audio file, from a European collection
The national collections are harder. The National Library has 100,640 radio programmes anyone may listen to, and each one already carries a IIIF manifest that names the recording and the rights. But the media server answers requests from nb.no only, so the file cannot be played from here. European collections often serve the file openly instead. Here is an example from a Europeana record, which says what kind of object it is, and offers a player to be embedded.
A video file, from a Norwegian museum
Film works the same way. DigitaltMuseum holds 2,298 records with a film, 1,166 of them under a Creative Commons licence. The picture in the record becomes the poster frame, so the reader sees what the film is before deciding to fetch it.
A video file, from a European collection
Here is the same thing from a European collection, for comparison.
Privacy
The audio and video players above are marked up with preload="none", so no audio or video is fetched before a user presses play. This also means that the collection’s server learns nothing about a reader who scrolls past, which preserves privacy for the user.
Licensing
All four recordings carry an open licence, which is why they can be played from here. The two Norwegian ones are CC BY and CC BY-SA, the European audio is CC BY-SA, and the European video is marked as public domain. The licence is read from the record each time the page loads, so it is the collection’s statement and not a copy of it.
How it works
A page says which object it wants, by source and identifier:
{% include heritage.html source="nb" id="URN:NBN:no-nb_digifoto_20180307_00007_NB_NS_NM_05512" %}
{% include heritage.html source="dimu" id="3df10c96-b33b-45c1-92bf-d9211ce574c8" %}
The identifier is the persistent one wherever the collection has one. For the National Library that is the URN:NBN, which the library undertakes to resolve for as long as the object exists, through urn.nb.no; the page looks it up to the catalogue’s item when it loads, and links to the URN. DigitaltMuseum’s object identifiers, Europeana’s record identifiers and KulturNav’s are already the stable ones those services cite.
The reader’s browser asks the collection for the object and shows what comes back. The rights line is the collection’s own statement, and the link leads to the object’s page at the source. Without JavaScript the link alone is shown.
What appears depends on what the record says the object is.
A picture comes as an address the browser loads directly. Where the collection also runs a IIIF image service, the include asks that service for the picture instead, and the reader gets a deep-zoom viewer, OpenSeadragon, served from this site. Nothing is copied here: the tiles come from the collection as the reader zooms.
An audio file or a video file comes the same way, as an address, and the include puts it in an HTML audio or video element with the browser’s own controls. There is no player library and no third-party embed code. The address is the one the record gives as its main file, so the bytes come from the collection’s own server when the reader presses play, and the collection can count that play in its own statistics. A video record also carries a still picture, which becomes the poster.
The same rule decides all three: the record says what kind of object this is, the include shows it accordingly, and the rights line under it is whatever the collection states.
Europeana and KulturNav are now in too: Europeana for collections across Europe, through its public demo key, and KulturNav for the authorities, people, places and terms, that objects refer to.
Linking data
The same object is often held in more than one place. The painting above is in the National Museum’s catalogue, in DigitaltMuseum, and in Wikidata; the fiddle in Bø is in its museum’s catalogue and in Europeana. Each of them gives it a different number.
Wikidata keeps those numbers next to each other. An item there can carry a DigitaltMuseum identifier or a KulturNav identifier as a property, so asking Wikidata “who has this identifier” gives back the item that stands for the same thing, if anyone has made the connection.
That is what the include does after it has shown an object. It asks Wikidata whether any item carries this collection’s identifier for it, and adds a link where exactly one item answers. The painting has such an item; the fiddle in Bø does not, and the line stays as the collection wrote it.
The value is not the extra link. It is that a page which mentions an object can reach everything else known about it, in any collection that has joined the same identifier, without anyone writing those connections by hand. Making more of them, and giving some back, is the subject of issue #50.