How to Contribute Data

In the spring of 2013, we took a new approach in establishing links from RIC types in OCRE to physical coins and hoards which reference those types. The model has evolved over time to be compliant to the new Nomisma ontology. The following models are required before contributing data into the Nomisma triplestore for use in OCRE, CRRO, and other coin type projects.

The first basic prerequisite for participation is that each object you wish to contribute be accessible to the web with a distinct URI. This is the core standard for Linked Open Data. After this, there are essentially two required elements for contributing data to a Nomisma-affiliated project, influenced by the requirements for joining the Pelagios Project: an RDF data dump (which also much be accessible by URL) and some metadata about this dump, in Vocabulary of Interlinked Datasets (VoID) RDF.


				<rdf:RDF xmlns:rdf="" 
	<void:Dataset rdf:about="">
		<dcterms:title>The Fralin | UVa Art Museum Numismatic Collection</dcterms:title>
		<dcterms:description>The Fralin Museum of Art at the University of Virginia numismatic 
			collection contains about 600 coins of mainly Greco-Roman origin.</dcterms:description>
		<dcterms:publisher>University of Virginia Library</dcterms:publisher>
		<dcterms:license rdf:resource=""/>
		<void:dataDump rdf:resource=""/>

The void:Dataset should include the URI for your project, and this URI should be linked from each individual object (physical coin, coin type, or hoard) within the data dump using the void:inDataset property (see examples, below). The dcterms:title, dcterms:description, dcterms:publisher, and dcterms:license are all required. The title and description must be textual strings, but the publisher and license may be URIs. In fact, the license should be a URI designating the dataset under an open license, such as those published by Creative Commons or Open Data Commons. A literal rights statement (dcterms:rights) may be made instead of a dcterms:license link.

Data Dump Model

The data dump model should conform to the standards established by, including the use of the Nomisma ontology where applicable.

Physical Coins

<rdf:RDF xmlns:dcterms=""       
	<nmo:NumismaticObject rdf:about="">
		<dcterms:title>Antoninianus of Gallienus, Rome, 254-255. 1991.17.140.</dcterms:title>
		<nmo:hasCollection rdf:resource=""/>
		<nmo:hasTypeSeriesItem rdf:resource=""/>
		<dcterms:replaces rdf:resource=""/>
		<nmo:hasAxis rdf:datatype="xsd:integer">6</nmo:hasAxis>
		<nmo:hasDiameter rdf:datatype="xsd:decimal">22</nmo:hasDiameter>
		<nmo:hasWeight rdf:datatype="xsd:decimal">2.86</nmo:hasWeight>
		<dcterms:isPartOf rdf:resource=""/>
		<nmo:hasObverse rdf:resource=""/>
		<nmo:hasReverse rdf:resource=""/>
		<void:inDataset rdf:resource=""/>
	<rdf:Description rdf:about="">
		<foaf:depiction rdf:resource=""/>
		<foaf:thumbnail rdf:resource=""/>
	<rdf:Description rdf:about="">
		<foaf:depiction rdf:resource=""/>
		<foaf:thumbnail rdf:resource=""/>

The title, publisher, and identifier are required, as is nmo:hasTypeSeriesItem. nmo:hasCollection is recommended for physical specimens. Before submitting your data to nomisma, please contact the administrators to create a URI for your collection. A physical specimen must be carry the nmo:NumismaticObject class. Finally, each object must contain a void:inDataset property linking to the URI of the void:Dataset in the VoID metadata RDF.

There are several fields which are particular to physical objects (and not hoards): nmo:hasAxis, nmo:hasDiameter, nmo:hasWeight, and obverse/reverse thumbnail and references images (to be added as necessary). Thumbnails and reference images that represent sides of coins should be contained in the data objects linked with the nmo:hasObverse and nmo:hasReverse properties. If they are not contained in the nmo:hasObverse or nmo:hasReverse, it is assumed that the images include both the obverse and reverse sides of the coin. None of these fields is required, though the measurements can be used for quantitative analyses. The measurements should be associated with xsd:decimal for weights and diameters, xsd:integer for axis (unless the source data uses fractional numbers for axes). The dcterms:isPartOf property is used to link a coin to a coin hoard that has a published URI.

IIIF Integration

The Nomisma RDF model has been extended to follow the Europeana Data Model specifications for defining IIIF images and services. Essentially, this requires additional triples about the reference image (foaf:depiction), which should be defined as an edm:WebResource. This edm:WebResource should link to a JSON-LD IIIF manifest or information object (dcterms:isReferencedBy) and the URI for the service (with the svcs:has_service property). The service should include several triples that defines it as a IIIF image service.

Note: The JSON-LD IIIF link in the edm:WebResource should be either:

  • If the image represents either the obverse or reverse, the link should point to the info.json for the image
  • If the image depicts both the obverse and reverse simultaneously (as is the standard for British Museum and Harvard Art Museums photographs), the link should be the full IIIF manifest.

<edm:WebResource rdf:about=";PTIF-1/full/600,/0/default.jpg">
    <svcs:has_service rdf:resource=";PTIF-1"/>
    <dcterms:isReferencedBy rdf:resource=";PTIF-1/info.json"/>
<svcs:Service rdf:about=";PTIF-1">
    <dcterms:conformsTo rdf:resource=""/>
    <doap:implements rdf:resource=""/>

Be sure to include the following XML namespaces/prefixes in your RDF:

PREFIX edm: <>
PREFIX svcs: <>
PREFIX doap: <>

Findspots for individual finds (not hoards)

Findspots can expressed with nmo:hasFindspot in the following way:

	<rdf:Description xmlns:geo="" rdf:about="">


<rdf:RDF xmlns:nmo="" 
    <nmo:Hoard rdf:about="">
        <dcterms:title xml:lang="en">Scurta (Romania; SCU)</dcterms:title>
        <nmo:hasFindspot rdf:resource=""/>
        <nmo:hasClosingDate rdf:datatype="">0014</nmo:hasClosingDate>
        <dcterms:tableOfContents rdf:resource=""/>
    <dcmitype:Collection rdf:about="">
        <nmo:hasTypeSeriesItem rdf:resource=""/>
        <nmo:hasTypeSeriesItem rdf:resource=""/>
        <nmo:hasTypeSeriesItem rdf:resource=""/>
        <nmo:hasTypeSeriesItem rdf:resource=""/>
        <nmo:hasTypeSeriesItem rdf:resource=""/>
        <nmo:hasTypeSeriesItem rdf:resource=""/>
        <nmo:hasTypeSeriesItem rdf:resource=""/>
        <nmo:hasTypeSeriesItem rdf:resource=""/>
        <nmo:hasTypeSeriesItem rdf:resource=""/>
        <nmo:hasTypeSeriesItem rdf:resource=""/>
    <geo:SpatialThing rdf:about="">
        <foaf:name>Valea Scurtă (Romania)</foaf:name>

Like physical coins, the dcterms:title, and dcterms:identifier are required, and the resource must carry the nmo:Hoard class. Similarly, the void:inDataset property is required in the nmo:Hoard object to link to the void:Dataset URI.

Presently, the hoard model is very simple and does not contain all the information stored in the database or NUDS XML record. The nmo:Hoard object should link to a dcmitype:Collection, which merely contains a list of associated TypeSeriesItems (linked with the nmo:hasTypeSeriesItem property). These records are meant to facilitate basic mapping and timelines in OCRE and CRRO. We anticipate enhancing the hoard model over the coming year.

The nmo:hasClosingDate is optional, but if it is a year, it should conform to the rdf:datatype, xsd:gYear (four digit number, padded with zeroes. Negative numbers represent B.C. dates).

The nmo:hasFindspot may be expressed, as with physical coins above, with an rdf:resource or the W3C geo ontology. The nmo:hasFindspot is not explicitly required, but without it, the hoard will not appear in maps in OCRE and CRRO. However, if a findspot is absent, and the nmo:hasClosingDate is present, the hoard will still appear in the timeline in the OCRE coin type record page.