Facilitator(s)
John Kunze (@jakkbl)
Abstract
The average URL breaks in 100 days, which is a minor disaster for institutions that care about durable access to the scholarly, scientific, and cultural record. It also fails to protect our collective investment in Linked Data and in AI. For responsible AI in particular, transparency and resilience are at risk when links (URLs) break to billions of inputs, and the already enormous financial and environmental costs of AI are amplified when links break to AI artifacts that then have to be reconstructed.
Citations that fail undermine all our preservation efforts, and it matters little whether linking to human- or AI-generated images, video, documents, datasets, software, maps, training materials, or intangible objects such as vocabulary terms, historical figures.
This one-hour tutorial introduces ARK (Archival Resource Key) persistent identifiers. As decentralized, non-paywalled PIDs (persistent identifiers, permalinks) for information objects of any kind, ARKs support web addresses that do not break (e.g., that do not return 404 Page Not Found). ARKs are similar to DOIs used in traditional publishing in that they both were introduced over 24 years ago, exist in large numbers (8.2 billion ARKs, 257 million DOIs), and support research and scholarship, appearing in the Data Citation Index, Wikipedia, ORCiD.org profiles, etc.
In contrast, ARKs are cheaper, more flexible, and less centralized, allowing providers to add any kind of metadata and to create identifiers in unlimited numbers. This is vital for inclusion of under-resourced institutions in the global South, where ARK adoption is accelerating. Long term commitment is never free of cost or effort, but no one pays for the right to assign ARKs, which can be especially useful for organizations (even those that are well-resourced) that need large numbers of PIDs.
Since 2001, billions of ARKs have been created by over 1600 organizations – publishers, libraries, data centers, archives, government agencies, and vendors. ARK strings also have compact forms that make them citation-friendly, as well as syntax rules supporting inferencing around containment (hierarchy) and variant forms. which makes them well-suited for AI and Linked Data applications.
No prior knowledge is required.
To register your participation in this workshop click on the “Going” button above. You will then receive an email notification as soon as facilitators post an update. Watch out to not register for two parallel workshops.