ARK Alliance

Home of the Archival Resource Key (ARK)

Arklet-frick: bringing ARKs from need to implementation at the Frick Collection

Exterior of the Frick Collection building in New York City (photo by Ajay Suresh, CC BY 2.0)

In 2021, the Frick Collection announced that it would start assigning ARK identifiers to assets in its digital collections and library catalog. The decision came after a migration to a new ILS (integrated library system) and the obsolescence of the old system’s identifiers and permalinks. The breakdown in both public links and systems interoperability spurred the Frick Collection to explore assigning persistent identifiers and ultimately to select ARKs.

From May to September 2023, the Frick Collection worked with a developer, Emery Infrastructure, to implement software that could mint, manage, and resolve ARKs. The software expanded on arklet, an ARK minting and resolving tool designed and maintained by the Internet Archive. The modified tool, arklet-frick, makes several improvements and changes to arklet.

In addition to supporting URL resolution, arklet-frick supports six Dublin Core metadata fields: title, identifier, source, relation, type, and format. Additional features supported include suffix passthrough, content negotiation for metadata delivery, and a command line interface. Future improvements will include support for subparts and commitment statements.

To create ARKs with arklet-frick, which uses a technical base of Python and Django, an organization must first request a NAAN (name assigning authority number) and enter it into the Django administration interface in order to generate an API key. This interface is also where administrators can create and manage shoulders, which subdivide a NAAN’s namespace. Shoulders are one of the most powerful and flexible features of the ARK specification because they allow ARKs to support a variety of content and usage standards. The Frick uses separate shoulders for the DAMS and the library catalog, which helps keep ARKs organized and easy to understand.

With a NAAN, shoulder, and API key in hand, users of arklet-frick can immediately start using the command line interface to experiment with ARKs. The application supports three key activities: minting, updating, and resolving. When minting ARKs, users can supply metadata for both the URL and Dublin Core fields. Resolving ARKs generically redirects users to the value stored in the URL field, but users can also append a “?info” or “?json” extension to see metadata directly. The Frick uses DigitalOcean as a provider for its production deployment of arklet-frick.

The Frick ran its testing phase for arklet-frick from September 2023 and to the end of the year. During this time, the Frick integrated the Django application into its existing systems, assigned user roles, created shoulders, and generated API keys for testing and production use cases. For each shoulder the Frick created, we created a set of rules governing the use case for the shoulder and the standards for the Dublin Core metadata fields.

The open source code for arklet-frick is available on Github, where we welcome feedback. In a separate blog post we will present how the Frick has been using ARKs since launching its API.

Recent Posts