People are the central nodes of the OML. We have created public profiles of composers, musicologists, performers, etc., effectively forming a research hub of people. But when someone signs up, that person too becomes part of this academic music network—scholars may find that a hub has already been formed with all of their publications, book reviews, and scores. They may choose to contribute preprints or new materials that might not have been already indexed or added to the network by another researcher. Collaboration networks also emerge from these various links, allowing researchers to identify, for example, when and in which works Erik Satie and Claude Debussy collaborated or influenced one another.
In addition to making scholarly content more accessible, such an interconnected platform promotes the exposure of related items from a variety of collections that might not have been easily discovered—for example, when manuscripts by a particular composer are spread across and held by multiple national libraries around the world. This benefits archives, but also librarians hoping to promote resources to which they devote their limited funds.
How were the participating archives recruited?
We conducted a wide research of digital collections relevant to the academic music community, which also included speaking with music librarians, researchers, professors, and students from multiple countries and schools. We shortlisted six national libraries with open platforms offering public domain resources, and that was our starting point for the beta phase of the project. Since then, we have added more collections—the Brazilian and Colombian national libraries, most recently—and we are working on new partnerships with other libraries and archives worldwide.
How do using shared ontologies, linked open data, and principles of the semantic web in the OML help users discover music?
Integrating and linking digital collections from a variety of sources require a tremendous amount of data alignment. Multiple languages, alternative spellings, role definition variations—to name a few—make it difficult to establish meaningful connections between disparate materials.
Alexander Street has managed this issue by building taxonomies over the past many years in order to publish its multimedia databases. The very same taxonomies that allowed us to build award-winning products and establish semantic links between materials (e.g., a video performance of an opera and its related scores) serve as the initial foundation of the OML.
With the launch of the OML, these controlled vocabularies have been made open so that others may align their collections to them and reduce the need to reinvent taxonomies with each newly created resource or each new digital collection published on the web. This means that every musical work indexed receives a public permanent identifier in the form of a URI (Universal Resource Identifier), facilitating representations of that musical work over the network. The opera La Bohème, for example, is located at the URI openmusiclibrary.org/work/1/la-boheme, which also serves as the starting point for a researcher to discover all resources associated with it.
Both the Spanish and the French national libraries have remarkable open data platforms that greatly facilitated metadata alignment and integration of their own resources with the wider OML network. They are available at datos.bne.es and data.bnf.fr, respectively. This was achieved through existing integrations between their catalogs and other open networks such as the VIAF (Virtual International Authority File) and Wikipedia.
Is there an effort to balance how much of each content type there is? How are the curators chosen, and how can interested parties get involved?
We started with articles, books, and scores, which we identified as the content types most widely used by music researchers, and then we shortlisted content providers based on what is most studied. Today, we have indexed 200,000-plus scores, of which 150,000-plus are OA. We have more than 1 million article citations, from 600-plus periodicals, and we have just launched videos with 2,000-plus titles from Alexander Street’s for-fee collection Music Online: Classical Performance in Video and thousands of other OA titles from YouTube and Vimeo, including full-length concerts in high-definition, opera performances, public lectures, TED talks, masterclasses, etc.
The OML is an open network, which means that everyone is not only able to access its index, but also to add new content to it. Interested parties can get involved by creating a free account on the site and contributing new citations, uploading articles and videos, curating lists of resources, and engaging in discussions with other music scholars in the OML forum.
What can organizations learn from the OML about creating their own open networks?
We have already developed very strong taxonomies for classical music, and we hope that others will make contributions that will cover other genres too (jazz, folk, electronic dance music, etc.), bridging the gap that currently exists between how resources are used and researched. Furthermore, additional genre coverage can potentially surface influences and connections that were not apparent to the music researcher before.
Engagement with the academic music community is critical to successfully growing and curating the OML. We have held panel sessions and demonstrations of the OML at the American Musicological Society (AMS), Music Library Association (MLA), and the International Association of Music Libraries, Archives, and Documentation Centres (IAML) annual conferences, where music librarians and scholars have provided feedback on platform design, functionality, relevance, and additional sources to include. This has enabled our team to successfully carry out agile product development techniques to arrive at the existing solution as well as ongoing developments.