Policies of the Penn Libraries
Open metadata benefits the libraries' mission:
Sharing, using, blending, and repurposing metadata improves the discoverability and usefulness of library resources. It also enables libraries to more effectively manage large collections. Therefore, the Penn Libraries commit to making metadata describing their resources as open and as reusable as possible.
Make our metadata as open as possible, and clearly label open metadata as such:
The Penn Libraries will not place restrictions on the use of metadata that it creates for information resources,* and will not place restrictions on other metadata beyond what is required by its creators. Open metadata will be clearly identified as such, to promote reuse. We recommend associating a Creative Commons CC0† declaration for metadata that we create, or that is otherwise in the public domain, to make its public domain status clear.
Seek open metadata sources, and honor any license conditions on republishing:
When we republish metadata from other sources, we will seek out open-licensed metadata where feasible, and publish the open license that applies to that metadata. (For example, Voyager catalog records that we create will be published with a CC0 annotation, and Voyager catalog records that are derived from OCLC WorldCat, which currently uses an Open Data Commons ODC-BY‡ license, will be published as ODC-BY with a credit to OCLC.)
Make our metadata formats easy to interpret and reuse:
Metadata will be published in accessible formats. We will publish metadata using well-documented standards where possible, and will also publish information about any formats or profiles particular to our use, also under open licenses. For instance, if we publish repository records as PQC-METS, we will also openly publish the PQC specification.
Make our metadata easy to access, both by people and machines:
Metadata will also be published in accessible protocols. We should make it as easy to get machine-parsable metadata as human-readable metadata for our resources, preferably through similar mechanisms. For example, a URL that shows a human-readable Franklin record should also be usable (possibly with a slight transformation or content negotiation) for delivering the same information in one or more machine-parsable formats, such as XML, an RDF linked data format, or microdata in a well-documented schema.
Support bulk download of metadata:
Where feasible, we will also support the downloading of metadata in bulk, so that interested users can analyze and reuse metadata at the full collection level. Ways to make metadata in bulk available include OAI-PMH servers (already available for some of our metadata), ResourceSync, or periodically generated dump files.
A partial list of metadata to open:
Metadata that we can openly publish includes catalog records in our Voyager database and Franklin catalog (including both bibliographic and holdings records), metadata for our digital repositories and other collections of digital objects, finding aids, and other bibliographic and authority databases.
Implementing the Penn Libraries open metadata policy:
*Confidential individual and business information out of scope: This metadata policy does not apply to metadata on individual library users, or to information about their use of library resources, that could be used to identify library users and monitor their uses of library resources, directly or through indirect inference, without their clear, informed consent. We follow library ethics and state law in preserving the privacy of individual library users and their use of library resources. Confidential business information is also not subject to this policy.
†CC0 1.0 Universal (summary): creativecommons.org/publicdomain/zero/1.0/
CC0 1.0 Universal (full text): creativecommons.org/publicdomain/zero/1.0/legalcode