Navigating the future of special collections metadata by using insights from the past 

During a conversation about our respective work projects last year, my colleague Chela Scott Weber mentioned a 2009 OCLC research report called The Metadata IS the Interface: Better Description for Better Discovery of Archives and Special Collections, Synthesized from User Studies. The report brought me back in time (much like a silver DeLorean) to my work as a graduate assistant in Louisiana State University Special Collections working on archival collections. The first collection I worked on there was the Louisiana Folklife Program project files, Ms. 4730, which is 44 linear feet, 21 video cassettes and 3,618 sound cassettes of primary source material about Louisiana folklore and the state agency devoted to preserving it. In 2002 when I worked on the collection, discovery was provided from both a record in the online catalog and a finding aid (written in a word processing program) available online as a PDF and in hard copy in the reading room. 

This collection and its discovery tools are excellent examples of what Jennifer Schaffner described in the report, so it was no surprise that it resonated with my experience in 2002. Schaffner focused on user interest in resource content and the gap between traditional metadata practices and user expectations. Although it was written several years ago, the report could easily be describing discovery needs for archival and special collections today if a few references to defunct products and programs were updated. In 2009, linked data was not a regular part of library and archival metadata discussions, but Schaffner’s points about a more user-centered approach to metadata apply equally well in a linked data context. As Chela noted in her 5 February 2025 Hanging Together post, this is a report that remains useful to our current work, so my post offers a synthesis of the points that resonated with me. 

The importance of aboutness 

Schaffner concludes that “some thirty years of user studies teach that Aboutness and relevance matter most for discovery of special collections.” This seems an obvious conclusion, but the reason it resonates with me today is the importance of providing subjects to enable discovery of related resources. Fundamentally, linked data is about providing relationships among resources on the Semantic Web. Archives and Special Collections Linked Data: Navigating between Notes and Nodes, a report from the OCLC Research Archives and Special Collections Linked Data Review Group, explains the importance of relationships in special collections: “Relationships are critical to special collections: we have built and invested in collections purposefully because of their relationship to one another and to a topical focus or subject matter.” The has subject relationship proves special collections users with the opportunity to discover new relationships among seemingly unrelated collections. As Schaffner notes, archival description has often focused on Ofness (e.g., genre and form) and collection provenance than the subject content of collections. A user may certainly want to explore relationships among agents who created the resources, or through a combination of creator and genre relationships, but without subjects, archival collections are likely to remain hidden to the researcher unfamiliar with an institution’s collections or a user interested in a subject regardless of resource format. Minimal-level descriptions may be necessary to balance staffing shortages with unprocessed collections, but some subject terms make a huge difference in discoverability among collections with devised titles like “James family papers” and “Postcards collection.” This premise is supported by 2023 OCLC research report Summary of Research: Findings from the Building a National Finding Aid Network Project: “In terms of what kinds of material might address these needs, more than half of pop-up survey respondents indicated that they are interested in any type of material relevant to their topic (55.8%).” 

Users expect results ranked by relevance 

This point certainly reflects an expectation that library discovery tools incorporate relevance rankings that are fairly common in internet searching. It also reflects one of S.R. Ranganathan’s five laws of library science—save the time of the reader. Imagine an archive has eleven collections with the subject heading “Suffragists.” When a researcher asked to see an archival collection, they may be presented with two book trucks full of boxes so saving time is critical in conducting research with primary source materials. A finding aid may narrow it down to a box, but a single box may contain 100 or more items. And then the researcher needs to repeat this process ten more times because there are eleven collections—that is a huge amount of time, and far more than would be spent looking at eleven books with indexes for information about suffragists. What relevance means for archival collections is a particularly tricky issue as many collections will have generic devised titles and related agents unique to a single collection. Thus, providing users discovery tools that enable them to filter results by multiple criteria including subjects, dates, language, and format, allows users to create their own relevance criteria with each search. 

Discovery tools are not obvious 

This is another conclusion that applies generally to libraries and archives, but especially to special collections resources that usually do not benefit from initial discovery in online bookstores and social cataloging websites like Goodreads. Users’ unfamiliarity with search types and indexes creates even more invisibility of special collections resources in online catalogs because of their uniqueness, e.g., archival collections titles may seldom benefit from indexing with keywords. An advantage of using archival linked data is the removal of such barriers to access. Eliabeth Russey Roke and Ruth Kitchin Tillman note this as one of the advantages to archival linked data. In their article “Pragmatic Principles for Archival Linked Data,” they write, “By making our data more accessible on the web using the semantic structures and language of linked data, we meet our users where they are, rather than expecting them to come to us.” 

Thus, digitization of special collections may greatly increase their discoverability, and here we have seen advancements since Schaffner’s report. 3D digitization of cultural heritage objects offers better possibilities for access to these resources virtually. We also have a relatively new tool, artificial intelligence, which offers the potential to transcribe and translate manuscripts more quickly. (I am now reminded that I should revisit the 2024 AI4Libraries presentation “Using AI Tools to Improve Access and Enhance Discovery Capabilities in Archival Collections” by Sonia Yaco, Digital Initiatives Librarian, Rutgers University Libraries.) 

Final thoughts 

Reading this report started with a personal reflection on my work as a graduate assistant, but it quickly turned into a realization of the importance of focusing on user needs as we plan for the future. Controlled vocabularies, so important for pre-coordinated searching, provide us with building blocks to create linked data with meaningful relationships among entities. Search engines and online catalogs save the time of users that use to flip through pages of paper finding aids and printed catalogs, but the amount of information available to users makes Ranganathan’s law “save the time of the reader” equally relevant today. Making my own addition to Ranganathan’s laws, I add, “use technology responsibly to save the time of the archivist and librarian.” Our metadata has been converted from catalog cards to online catalogs to linked data, but fundamentally the why doesn’t change. It’s all about the user.