Updating search context in RRDP: coin types as a search facet

I have finally had the time to enhance the search indexing of dies in the Roman Republican Die Project. Although RRDP dies may have the RRC number in the ID portion of the URI or in the title of the record, there is no explicit link in RRDP metadata from a die to a type. Instead, dies are linked to specimens, and specimens are linked to types. The initial version of RRDP only indexed explicit metadata into Apache Solr for the search and browse page--in this case, control marks represented by symbols, letters, or monograms. The die -> specimen -> type graph can only be extracted via Nomisma.org's SPARQL endpoint, and I have just recently implemented this SPARQL lookup into the Solr indexing process by creating an API in Numishare to extract a list of coin types associated with a die URI.

 

Updated RRDP browse page with coin type facet

As a result, the coin type(s) are now visible in the browse page and can be used for narrowing down the query response. In addition, the list of associated types is visible in the RRDP record page, below the typological description and above the die example photographs.

Updated layout for subtypes in Numishare

Upon request for Coinage of the Roman Republic Online, the layout of subtypes for a parent type has been modified to present the subtypes in a table format with example images presented much like the browse page or types associated with Nomisma concepts (e.g., Seleucus VI). The table includes columns for obverse and reverse (where most differences between subtypes are apparent). This format makes it much clearer which elements of the typology vary, for example differences in legend or symbols on the obverse and/or reverse of the coin.

RRC 363/1 and subtypes.

The "Examples of this Type" section has been moved down under the subtype list. The example list still contains all specimens linked to the parent type or any level of subtype below this, and can be paged through by clicking the relevant buttons in the interface. If you want to see only specimens associated with a subtype, click on the link in the subtype label to view the Numishare record page for that subtype, and all example specimens associated with the subtype will be available for view or download.

Seleucid Coins 1.1 and subtypes, with variations in symbols

The Egypt Centre at Swansea University joins OCRE

The Egypt Centre of Swansea University is the latest partner in the Nomisma.org consortium. The museum houses 6,000 objects from Egypt and the ancient world, including Greece, Rome, Cyprus, Mesopotamia, and Nubia. Among these are about 50 Roman imperial coins, which have now been integrated into Online Coins of the Roman Empire.

RIC II Nerva 30 (denarius)

Although the collection is small, it does provide the sole photographic example of at least one type, Nerva 30 (denarius).

New version of Nomisma.org site released

Last Friday was the soft launch of the new Nomisma.org website. The design aspects of dynamically generated pages are largely unchanged (search results based on Solr, interfaces built on SPARQL, concept pages generated from RDF), but we have migrated all static content to Jekyll. This will make it easier for Nomisma committee members to update pages through Github, making the project more sustainable in the long-term since I (Ethan Gruber) won't have to update these static pages within the Nomisma framework.

 

The new site introduces an improved structure for getting information about Nomisma. It includes:

• An about page with the history of the project and list of current projects.
• The scientific committee.
• A list of Nomisma working groups and whom to email about contributing.
• A landing page with more concise information about how to get Nomisma data.
• Improved UI for SPARQL examples (expect more examples to come) and two YouTube videos that introduce Nomisma's data model and SPARQL endpoint.
• A link to a Digital Numismatics bibliography (a Zotero group) maintained by the scientific committee
• Furthermore, I made an improvement to the partner datasets page in order to categorize contributions by dataset type (collections, types, hoards, etc.).

 In the near future, you should expect the publication of the Nomisma cookbook. This cookbook is our long-overdue detailed documentation of our ontology and data model, with specific modelling examples represented as TTL.

New version of Coinage of the Roman Republic Online released

Today, an improved version of Coinage of the Roman Republic Online (CRRO) was released. This release improves the consistency of the data entry and implements the type-subtype hierarchical structure in Crawford's original RRC typology. The CRRO types are no longer a flat structure of about 2200 entries, but there are now about 1750 parent types and about 500 subtypes, with subtypes links to parents through the skos:broader RDF property upon ingestion into Nomisma.org. Improvements were made in other areas as well:

• Issuers represented by symbols or letters were generated as Nomisma.org URIs. Some symbols were reused in different periods, implying different moneyers, and so each one has a different URI, e.g., http://nomisma.org/id/star_1_issuer_rrc.
• Type descriptions and legends were cleaned and normalized. Control marks of various types were split into different positions with respect to the obverse and reverse of the coin. These symbols are now queryable in the browse page, much like Hellenistic Royal Coinages.
• Deities URIs from the British Museum were replaced with new Nomisma.org ones. Several new deities were created in Nomisma that were not accessible in the search facets before, although we certainly have not isolated all of them yet.
  
• People who appear on the coins are now searchable through a portrait facet. These types mostly affect coins of the First and Second Triumvirates, although some miscellaneous regnal or early Republican period people were created in Nomisma.org.
• The mint facet was cleaned up so that only true Nomisma.org mints appear in the list (or "Uncertain Value") and historical regions moved into the hierarchical region facet. We rely on Nomisma's geographic hierarchy to populate the region facet when we are able to identify the mint.
• All of the Roman Republican issuers have been linked to the Nomisma concept for the Roman Republic, with start and end dates for their role(s) as a moneyer.
  
• We adopted the EpiDoc TEI specification for ligatures in the legends (see this query result and related types).

 

Example of browse page with ligatures in legends and queryable symbols.

 

Background

When CRRO was published in 2015, it had been derived from cataloging data from the British Museum database, curated by Eleanor Ghey. Specimens that were not represented in the BM were supplemented by data entered by Rick Witschonke. At the time, Rick was entering the late stages in his fight against cancer, and very much wanted to see an "OCRE" for Roman Republican Coinage. We expedited the publication of CRRO as a result, even though it was not quite as robust at the outset as OCRE was. Despite the popularity of CRRO (second only to OCRE in usage out of the ANS' digital projects), we have not been able to revisit the data structure of the project in over 7 years. Furthermore, the original Excel spreadsheet from which the CRRO data originated no longer seems to exist or be accessible online.

Several months ago, I processed the NUDS/XML files from CRRO back into a spreadsheet that I began to normalize in OpenRefine. After some clean-up, this was uploaded as a Google Spreadsheet for more fine-tuned editing by the ANS curatorial associate, Alice Sharpless (who primarily works on the Roman Republican Die Project). Alice did a significant amount of normalizing legends and type descriptions, fixed issues with the conflation of region URIs as mints, and split symbols into multiple columns.

Next Steps

Now that CRRO exists as a canonical spreadsheet that can easily be edited or updated, we plan to insert some new Republican typologies that have been published since RRC in 1974. These new types can be slotted into the spreadsheet as necessary with an appropriate bibliographic reference. These new types are necessary for further cataloging both in RRDP and Coin Hoards of the Roman Republic, which will hopefully see a new edition from Kris Lockyear.

Nomisma and ANS digital projects accessible at HTTPS

Following the migration of the American Numismatic Society website and digital projects (including Nomisma.org) from our old Rackspace server to a new cloud server hosted by Amazon Web Services, the ANS website (Wordpress) has been fully migrated to use an SSL certificate and HTTPS. The same certificate has been applied to the other ANS digital projects to enable secure interactions via HTTPS URLs, but the old HTTP URLs do not forward automatically to HTTPS. The main reason for this is that http:// and https:// URIs are considered difference in Semantic Web applications, and migrating to HTTPS creates a number of unpredictable downstream effects for consumers of our machine readable data.

Those who use SSL in their own information systems and want to consume Nomisma or ANS machine readable data using Javascript in a web browser will be able to do so by replacing http with https in the web service URL.

The ANS IIIF image server has incorporated automatic forwarding to HTTPS, and so the advantage in this case is that other systems that use HTTPS (such as the Digital Library of the Middle East, Peripleo, Iron Age Coins in Britain, and other external aggregators) will be able to securely load JSON resources from IIIF manifests and the image server.

Updates to mapping functionality in Numishare

At long last, I have implemented some contextual changes in the display of points in the maps pages in the Numishare platform, particularly with respect to the sizing of points to reflect density of distribution (like what had already been implemented in maps on nomisma.org and record pages in Numishare).

The changes can be summarized as follows:

• Mint, findspot, and geographic subject (applies to Art of Devastation only, so far) layers have variable sizing based on density.
• The marker cluster plugin has been disabled for individual findspots, so each point is readily apparent without further zooming into the map.
• The findspot layer now refreshes correctly after narrowing the search facets down.
• Symbol facet menus now display the correct, human-readable label, and monogram SVG graphics will appear in the menu, as it functions on the browse page.

These visualization updates have also been implemented in the map popup window on the browse page.

The changes are most readily apparent in the Celtic Coins Index Digital. 

Findspot distribution of the Corieltavi.

 And Hellenistic Royal Coinages clearly illustrates the highest level of production in Amphipolis and Alexandria:

And while the general map of the Roman Empire isn't particularly illustrative (about half of all types minted in Rome), when you drill down into specific queries, the variations in production are more apparent.

Mint distribution of Constantine I