Provenance and Arrangement

Archival decisions about how to describe a collection's provenance are sometimes not as straightforward as they might seem. Jane Zhang points out, in her discussion of the origins of archival theory from its European beginnings in the nineteenth century, that the “principle of provenance mandates that records of the same origin should be processed as one record group and not be intermingled with records of other origins when accessioned, organized, and described into archival collections.”² However, the first edition (2004) of Describing Archives: A Content Standard (DACS), in Part II's “Introduction to Describing Creators,” indicates:

The structure and content of archival materials cannot be completely understood without some knowledge of the context in which they were created. It is insufficient for the archivist simply to include the name of the creator in the title of the description of the materials. Additional information is required regarding the persons, families, and corporate bodies responsible for the creation, assembly, accumulation, and/or maintenance and use of the archival materials being described.³ [Emphasis added.]

In deciding who a collection's creator is—and what its “origin” is—archivists sometimes have to go beyond simply describing who wrote a letter, for example, to give the full picture and context of how that letter relates to the materials around it and why it is in a repository in the first place.

Subject-based Collections

The problem with a creator-based arrangement was that the archivists knew very little about the creators of many of the materials, particularly the correspondence. Records for the small Wade Hall collections only denoted the creator by first name and a location. Rearranging these collections with the donor, Wade Hall, as the compiler still aligns with the concept of provenance, but the additional step to arrange the disparate materials into subject-based collections starts to violate the idea of respect des fonds, which would seem to require keeping them separate to preserve context. The purpose of making these collections subject-based is to create context artificially to help the user discover relevant materials. Authors began exploring this tension of creating context without violating the principles of provenance starting in the 1980s. Richard Lytle studied how users approach research in archives by examining subject queries with provenance (Lytle's term for linking subject queries to the administrative histories and biographies of a collection) and context indexing (Lytle's term for matching subject queries with an index or a catalog). Lytle's study found that users tend to seek archival collections on a specific subject, and it exposed the flaws in a provenance-based method of arrangement, as users depend on an archivist to locate collections of interest on a specific subject.⁴ In 1982, Mary Jo Pugh criticized strict adherence to provenance and original order and argued that users want to locate archival materials by specific subjects. At the time Pugh was writing, this was only possible through the publication of a specific subject guide on a topic of interest or through an in-person reference interview between the user and the archivist.⁵ Building upon the work of Lytle and Pugh, Elsie Freeman argued that the archival profession was not oriented to users, was unaware of its user base, was uninformed of how research happens within the archives, and provided inadequate help. She advocated that archivists should perform usability studies.⁶ At the time Freeman, Lytle, and Pugh were writing, users visited the archives in person to view finding aids and materials, and they asked reference questions by mail or by phone, and the usability studies of the next decade reflect this environment. It would take time and technological innovation for the idea of a subject-based arrangement to percolate again to the forefront of archival literature.

Online Finding Aids and Digital Collections

The Internet revolutionized the archival profession through the creation of online finding aids and digital collections. In 1997, Thomas Ruller predicted a world where most researchers would interact with archives online, leading to a reimagining of the finding aid and the rise of usability studies in the online environment.⁷ By the 2000s, Wendy Duff, Heather MacNeil, Max Evans, and many others were documenting the archival profession's varied responses to the online environment and its effects on description, usability, and providing reference services.⁸ The transformation of the finding aid to Encoded Archival Description (EAD) was another innovation.⁹ Corey Nimer and J. Gordon Daines III concentrated their usability study on online finding aids, suggesting that the age of search engines had altered users' searching behavior.¹⁰ Elizabeth Yakel and Deborah A. Torres conducted interviews with users of archives and found them unfamiliar with archival terminology, the structure of finding aids, and search strategies to locate information effectively. Helen Tibbo, Wendy Duff, Catherine Johnson, Donghee Sinn, and Nicholas Soares focused on historians and genealogists, showing the evolution in researchers' information-seeking behaviors while using archives.¹¹ With the rise of online finding aids and digital collections, users did not necessarily have to visit the reference room to interact with the materials and seek help from the archivist.¹² Elizabeth Yakel suggested this shift might change the way archivists keep information “within in its original context or supply a context that enables use in a contextualized way.”¹³ The new subject-based arrangement of the Wade Hall collections fits Yakel's idea of a contextualized arrangement. Jennifer Schaffner argued that users often rely on search engines as a means of discovery and search by subjects and key words. Archivists organize collections by provenance, but users are often interested in the “aboutness” of a collection instead of what it is made of (“ofness”). Schaffner concluded that professionals should close this gap between archival descriptive practices and the expectations of users, especially when dealing with minimally described collections.¹⁴

Case Studies—Physical and Digital Consolidation

Although the literature has advocated for archivists to incorporate the needs of users and the role of an online environment into archival description, it offers few practical examples to show how institutions have implemented these ideas. In 2003, Pam Hackbart-Dean and Sammie L. Morris provided reprocessing case studies of the ever-growing collections of Amelia Earhart at Purdue University and the International Association of Machinists and Aerospace Workers (IAM) at Georgia State University, which touch on many of the issues involved in the Wade Hall consolidation. They outlined four arrangement options for growing collections and the positive and negative outcomes of each. The reprocessing of the Earhart collection mirrors the first option of archivists physically and intellectually integrating additions to expanding collections, while the reprocessing of the IAM collection demonstrates the second option, where archivists intellectually integrate additions within a finding aid but do not physically integrate the material. Another option is to keep new additions physically separate and only add descriptions to the appendix of the finding aid. The final option is to treat additions as separate collections, with each collection having its own collection number, finding aid, and physical containers. That is what had happened in the case of the small Wade Hall manuscript collections. Hackbart-Dean and Morris found that understaffed institutions used this arrangement most often, but identified this approach as the least user-friendly, as researchers must examine multiple collections and disparate finding aids to locate records of a single creator. Although the four options presented are extremely useful for institutions considering physical reprocessing projects, Hackbart-Dean and Morris skirted the issue of how reprocessing affects previously digitized collections, only mentioning the lack of location information within the finding aid and metadata of digitized items and not going any further with the discussion. The Wade Hall Consolidation Project seeks to fill this gap by addressing how institutions might deal with existing location information in the physical and digital environment as well as larger issues of digital preservation.¹⁵

The University of Alberta's Prairie Provinces Collection is similar in scope to the collections of materials donated by Wade Hall, as it also includes printed, photographic, and manuscript materials that continued to arrive as both individual pieces and small batches over the years. While the previously processed Wade Hall small manuscript collections had legacy descriptions, the staff at the University of Alberta had not accessioned or described their Prairie Provinces ephemera, which allowed them to choose LibGuides as the organizational scheme for the finding aid. The staff initially arranged materials into broad categories (i.e., Photographs, Letters, Memoirs, and Posters), but found a need for additional categories and subcategories to describe this collection adequately. Ultimately, LibGuides proved to be an unwieldy platform, and the staff migrated the content to a traditional finding aid. The University of Alberta's experiences informed the University of Alabama archivists' decision to create separate subject-based collections instead of an overarching Wade Hall manuscript collection with multiple subject-based collections within it.¹⁶

Dickinson College houses one- to two-item collections in filing cabinets, with the only access point being a subject-based card catalog in the reading room. The staff adapted a blog detailing reference requests with tags to create a “catablog” for individual items with description, metadata, and digitized images to create awareness of these collections online and to allow search engines to index them. Web analytics and external linking to the blog from outside web-sites demonstrate the blog's popularity, and, in turn, its popularity leads to increased reference requests and research visits. Seeing how the tags (e.g., Civil War) within the blog produced artificially created subject-based collections, the staff created separate subject-based blogs, which drew the attention of a noted historian who found them through a search engine. Through search engine indexing, the file cabinets are discoverable.¹⁷ As Dickinson College's example proves, greater discoverability of these small Wade Hall manuscript collections is possible through enhanced description, rearrangement into subject-based collections, and metadata harvesting and search engine indexing of ArchivesSpace and CONTENTdm.

One of the most relevant case studies examines a project that incorporated legacy data in the creation of a subject-based digital collection at the National Agricultural Library. Christian James and Ricardo L. Puzalan's definition of “legacy data,” in which existing data is reused and transformed to create a new collection of data, applies to the way our Wade Hall consolidation team took legacy descriptions and metadata and repurposed them to create new finding aids and metadata for the subject-based Wade Hall manuscript collections. In their 2015 study, James and Puzalan concluded that the inconsistency of the subject terms, different vocabularies, and application of standards in legacy data diminishes the ability of users to search effectively and that contextualization of a subject-based collection allows researchers to make interpretational connections when items are placed side by side instead of organized as disparate collections.¹⁸

Case Studies—Legacy Data

The majority of the literature addresses legacy data in connection with a library's adoption of or migration to a specific management system, whether ArchivesSpace, a new digital asset management system (DAMS), or something else.¹⁹ Although these systems may be describing the same collection, an archival collection's finding aid often does so through folder-level description, while item-level metadata is typically the standard for digital collections. In 2013, Jane Zhang and Dayne Mauney sampled the depiction of both archival and digital collections online and identified the embedded, segregated, and parallel models of how institutions represent archival and digital collections.²⁰ This interplay between systems that manage physical collections (i.e., ArchivesSpace) and digital collections (i.e., CONTENTdm) that Zhang and Mauney identified is an issue specifically addressed in the Wade Hall Consolidation Project, but rarely mentioned in discussions of legacy data within the literature. Digital preservation of legacy metadata receives little discussion. In 2014, Todd Bruns, Stacey Knight-Davis, Ellen Corrigan, and Steve Brantley alluded to the challenges of dealing with legacy data and files, including digital preservation concerns, as their institution migrated digital content to a new institutional repository.²¹ This case study of the Wade Hall Consolidation Project addresses legacy metadata and the difficulties inherent with existing digital preservation practices by giving examples of how one institution dealt with this problem during physical and digital consolidation. Jane Darcovich, Kate Flynn, and Mingyan Li remediated legacy metadata from existing digital asset management systems in preparation for a migration as well as inclusion in the Digital Public Library of America (DPLA), a leading example of a multi-institutional digital content aggregator.²² The necessity of maintaining the older uniform resource locators (URLs) for materials already in aggregated sites is another consideration during migration. Persistent URLs (PURLs) are powerful and necessary tools in a shifting digital environment.²³ The conscious choice to leave metadata as-is for the consolidated Wade Hall digital collections that already exist in multi-institutional digital content aggregators, and rely on the persistent URLs to point users to the new digital asset management system, represents a contribution to the existing literature and a possible strategy to save institutional resources and time.

Personnel

In 2017, the Special Collections department hired a new associate dean who was tasked with supporting research and enhancing the reputation of the unit. This meant hiring two additional processing archivists, restructuring and filling the open special collections and digital initiatives librarian position, and changing the digital asset management system. These additions gave the unit enough resources to take on the consolidation project. In hiring new staff for its Special Collections department, university libraries directly confronted the issue of bringing the university in line with the evolution of professional standards for archivists. Except for one professionally trained archivist, all the other staff had transitioned from archives-adjacent areas, such as library cataloging, or learned everything on the job.²⁴ The newest hires were recent graduates from archival programs where they had been trained on the most current standards, descriptive terminology, and technology used in archives. Moreover, the newer processing archivists had also been trained in archival theory and were aware of an emerging trend in subject categorization, which they were able to bring to the job to enhance what had already been done—all the while stepping lightly so as not to criticize the work of the archivists who had come before them.

Migration to a Different Digital Asset Management System

The University of Alabama Libraries developed an open-source digital asset management system, Acumen, in 2010. Over the years, Digital Services staff digitized 18TB of content, including numerous small Wade Hall collections, and placed it in Acumen. In October 2018, University Libraries selected the cloud-hosted version of CONTENTdm as its new system and publicly launched a small selection of migrated digital collections in February 2019. The decision to move to a vendor-based system arose from several factors, including cost, level of technological development needed, the loss of the original personnel who developed and maintained Acumen, and shifts toward research and accessibility in the library's strategic plan. The timing of the Wade Hall Consolidation Project meant that Digital Services staff could digitally consolidate collections and migrate content at the same time. The digital consolidation also addresses the issue of having too many existing digital collections to fit into the finite number of digital collections CONTENTdm's mobile responsive interface could accommodate without sacrificing appearance and usability. In 2018, CONTENTdm's mobile responsive interface could only technically support 400 digital collections, which was problematic because the University of Alabama had 600 digital collections slated for migration, of which 287 were Wade Hall manuscript collections.²⁵ Finally, the mirroring of the analog and digital collections, with their subject arrangement, will enhance discoverability in an online environment.

Users and Search Strategies

In support of the university's goal to become a top research institution, the libraries' strategic plan called for the development of innovative library instruction and services for varied users, provided through multiple channels. To fulfill this objective in the strategic plan, Special Collections needed to identify its users and their search strategies before it could determine the best way to alter its current research. As the University of Alabama is a public university, users included community members doing local history or genealogical research, visiting scholars, graduate students, and administrators researching university-specific records. The majority of physical visitors are undergraduate students. Although analyzing exact user statistics on each of the small Wade Hall collections is outside the scope of this article, general experience shows that users searching for collections of the type that Wade Hall's materials represent—that is, collections that are not about a prominent individual or family—tend to search by subject. All the Special Collections staff who assist users in the reading room have had encounters with students who ask for resources about “civil rights” or “slavery” or “women at the university,” for example. Even before the widespread use of the Internet, Mary Jo Pugh pointed out that subject knowledge about records could not simply be located within the mind of a reference archivist but should exist somewhere attached to the record so that a user might be able to access that knowledge independently.²⁶ Today, patrons often investigate a repository's resources online before they make a physical visit—which gives Pugh's idea even more importance and validates the reasons behind the Wade Hall Consolidation Project.

Personnel

Under the leadership of Special Collections' associate dean, the Wade Hall Consolidation Project began with a team consisting of three processing archivists, the reference services and outreach coordinator, the archival access coordinator, and two staff members in Digital Services, the segment of Special Collections that oversees the digital asset management system. The newly hired special collections and digital initiatives librarian joined the team a few months into the project and serves as the liaison between the archivists and Digital Services.

Project Timeline

Special Collections' associate dean provided the team with a two-year time-line to consolidate the items that Hall donated into broader collections organized by category, with the disparate finding aids simplified into one record for each consolidated unit. Digital consolidation would follow once physical consolidation was complete for each collection. At the first meeting, the team established two phases of the project: the planning and preparation phase and the consolidation phase. The planning and preparation phase took approximately five months. The reference services and outreach coordinator identified the collections that would require consolidation (two months), and the archivists created a workflow and sorted the data into subject-based collections (three months). Once the planning and preparation were complete, the archivists embarked on the consolidation phase, spending 15 percent of their time (six hours per week) on the project for nineteen months. Digital Services spent one month creating metadata revisions templates and adopting existing migration procedures and migrated its first consolidated collection eleven months into the two-year timeline, with a new consolidated collection migrated each month after that point.

Criteria and Survey of Wade Hall Collections

The first step was to identify the relevant Wade Hall collections that needed consolidation, including the troublesome one- and two-item collections as well as other collections under one linear foot. Using a native bulk export of JSON-formatted data from ArchivesSpace, Special Collections' reference services and outreach coordinator isolated Wade Hall collections and cleaned the data up to make them useful for analysis. ArchivesSpace's Extent field provided the size of each collection in linear feet, but this measurement did not adequately represent the number of items in small collections. Using the manuscript number, she searched Acumen for an item count for existing digitized items and used the finding aids to determine the collections' true size. She created a size column and entered the number of items, or used “M” (for “many”) to designate more than five items in a collection. To aid the team in developing categorical themes, she scraped the JSON data to pull dates and subjects, and she began sorting collections. Through this process, she identified 1,429 physical collections, 287 of which were already digitized.

Using the JSON data, she identified groups based on the collections' subject headings, such as African American, Civil War, and Travel. For those collections lacking subject headings, she identified categories, such as diaries, scrapbooks, and business and finance, and grouped similar collections together. Finally, she took this data and organized them into an Excel spreadsheet with three section tabs: “Just 1–2 letters” collections; “Just 1–2 other” collections that are not letters; and “The rest,” which contains collections larger than two items or an unknown number of items (see Figure 1). Although the data may not have included all of the small, consolidation-worthy Wade Hall collections, they became a starting point for the processing archivists, and the spreadsheet served as the master for the project.

View Figure

• Download as Powerpoint
• Download Figure

Physical Consolidation Planning

Such a project involves considerable staff time and resources; the small creator-described collections that already existed were still findable but did not reflect Wade Hall's intent as the collector. Previous processors involved in creating records for the Wade Hall materials had named all of these small collections after their creators (the “origin”), as in “Samuel Holt Letter” for a single letter written by Holt, with its own resource record and collection number in ArchivesSpace. Upon reexamination of existing processed Wade Hall collections, archivists noted that Wade Hall had often written notes in pencil on the materials he had donated saying things like “Prohibition” or “World War II”—indicating that he intended them to be assembled according to a certain category. The team began to see that providing context to these records meant looking at provenance differently. Instead of seeing the provenance of the Samuel Holt Letter as coming from Holt, a man with no relation to the university and who did not convey his records to the archives here, they saw it as a piece of American life given to us by Wade Hall—Hall's intentions became the context that gave meaning to the record, showing it as one of many letters describing American southern business transactions. In other words, the earlier processors were not wrong to categorize materials by the creator the way they did, but it could be said that they did not provide the full picture of the context that gives the records meaning and shows how they are assembled.

The previous arrangement of these discrete Wade Hall collections made it difficult for researchers to discover relevant resources in ArchivesSpace. For example, before the consolidation project began, Special Collections held over a hundred small collections from the World War II era. Now these collections all exist as one subject-based unit—the Wade Hall Collection of World War II Materials—that makes it possible for researchers to find all the items within ArchivesSpace or CONTENTdm through a single search. With the consolidation, people can better gauge the volume of archival holdings on a particular subject, and it is easier for library staff to direct researchers toward other collections that might relate to their topics of interest. For instance, the newly consolidated Wade Hall Collection of World War I Materials complements Special Collections' separate, non–Wade Hall collection of World War I Posters.

One of the archivists' biggest concerns was to preserve necessary information from the original finding aid data. To accomplish this, they added general notes about the old collection numbers and titles to the consolidated collection's finding aid. This would allow researchers who had been familiar with the old collections to search with the same terms they used before and still find the collections they need within their new settings. The team added primary subject headings common among the old collections to the collection-level metadata of each new grouping and left any collection-specific subject headings at the item level. The archivists then streamlined any substantive information in the Abstract, Biographical/Historical note, and Scope and Contents note into a single file-level note in the new metadata. Reexamining the original finding aid data manually did take time, but proved valuable, allowing the archivists to reassess the previous wording and make judgments about its accuracy by comparing the information in the finding aid to the item in-hand.

Unclear Provenance

Over fifty years ago, Barbara Kaiser profiled donors, like Wade Hall, who donate in varying amounts and over a long period of time. She posited that the type of relationship initially established between the repository and the donor will determine “to a considerable extent the number and nature of ensuing problems.” Kaiser advised repositories to communicate their policies to donors early in the donation process.²⁷ Upon first establishing a relationship with a donor, an institution should clarify, among other things, how soon a donated collection is likely to be processed, whether the institution will discard unwanted materials, what sort of storage it can provide, and the kinds of restrictions it will apply to the donated materials. Some of the problems Special Collections archivists had as they tried to reorganize Wade Hall's donated materials stemmed from earlier curators not clarifying these policies at the time of intake—particularly in the area of restrictions on sensitive materials.

Wade Hall generally purchased his collections from estate sales, flea markets, auctions, rare materials dealers, and other sources. He rarely had any direct connection with the creators of these items—which he acquired largely in Kentucky and the lower Midwest, although the materials came from all over the United States and beyond. Some of the letters in the collection are relatively recent items from the 1980s and 1990s that contain private information from writers who had no idea their discarded correspondence would find its way into a public archival collection at the University of Alabama. While University Libraries has deeds of gift for the materials from Hall, Special Collections also realizes that the murky provenance of these items requires a unique approach to the repository's usual restrictions practice. To respect the privacy of the creators and their family members, Special Collections established a policy that restricts access to sensitive Wade Hall materials for seventy-five years from the date of creation. Previous archivists had restricted some items, though no consistent policy existed on how long the restrictions would apply. Throughout the consolidation process, Special Collections staff reexamined previously restricted Wade Hall materials to determine if the restrictions should be retained or if they could now be lifted.

Legacy Description: Inaccuracies and Assumptions

The team saw many inconsistencies in the ArchivesSpace records, but time had never been allocated to correct them. Such inconsistencies came about when staff imported records into ArchivesSpace from Archivist's Toolkit, or they existed because information input by earlier staff and students lacked uniformity and standardization, and because processing procedures had changed over the years. Because almost every item was reprocessed, the archivists corrected inaccuracies and assumptions found in the old finding aids. For example, some past finding aid authors hypothesized, without verification, that items were produced in certain locales. Archivists removed such speculations, allowing researchers to determine the locale by examining the evidence on their own. They also removed editorial commentary such as “the handwriting is unintelligible,” “poor grammar,” “spelling is atrocious,” or “the language of the letter could not be determined,” and substituted neutral language throughout the description. One such example was the Phillip Thompson Letter (see Figure 2). The processing archivist made several changes to its Scope and Contents note, such as removing gendered references to the unidentified letter writer (i.e., “his” children).

View Figure

• Download as Powerpoint
• Download Figure

The team also used this opportunity to identify the language of materials wherever possible and to correct typos, misspelled names, wrong dates, and other mistakes in the legacy descriptive metadata. This consolidation project was a chance to revisit every item and make sure it is described according to current archival standards.

Legacy Description: Local Subjects

In the past, Special Collections created its own nonstandardized local broad subject headings, such as “Architecture and Landscape” and “Civil Rights and Human Rights,” which staff members sometimes used in their finding aids instead of authorized Library of Congress subject headings. The consolidation project archivists, upon finding a suitable substitute, deleted these local subject headings and replaced them with authorized Library of Congress subject headings as new finding aids were created in ArchivesSpace for the consolidated collections. By standardizing subject headings and names, the archivists intended to make it easier for future researchers to locate collections similar to one another.

“Daily Life and Family” was the most problematic catch-all local subject heading as it appeared in more than 600 of these collections, and it was often the only subject heading. The content of these collections did not fit into a clearly defined category. The overwhelming majority of the “Daily Life and Family” materials were letters in which ordinary Americans discussed their everyday life in a certain place and time. The solution was to create a new umbrella collection called the Wade Hall Collection on American Life, dividing it into geographical series (i.e., Northeast Region, Midwest Region) based on the creator's location and arranging it in chronological order. This new arrangement provides users with examples of American life in a particular geographic region or during a specific time and allows for easier cross-comparison.

Processing Plan/Workflow

Once the archivists had time to go over the master spreadsheet produced by Special Collections' outreach coordinator and the accompanying notes, the team created two working documents for brainstorming ideas—one focused on workflow and the other on subject categories. Using the workflow document, the team proposed different approaches to this large project, outlined the responsibilities of each team member, and discussed potential rearrangement problems. After identifying frequently occurring categories from the master spreadsheet, the team recorded them in the second document and culled it to create the official list for the subject-based, consolidated collections. The team identified fourteen categories (e.g., Civil War, Diaries and Scrapbooks, and Travel and Tourism) that would form the consolidated collections. These categories came from both the assigned subject headings associated with the original small collections and the categories identified by Special Collections' outreach coordinator during the creation of the master spreadsheet.

The two brainstorming documents enabled the team to establish an official workflow, which consisted of two phases: planning and consolidation. The planning phase involved the redistribution of collection data into the new subject-based collections. The team created a new spreadsheet for each subject-based collection with tabs that aligned with the master spreadsheet, plus an additional tab for restricted items, and moved each original collection to one of the subject spreadsheets using the provided metadata.

After creating new subject-based collections, the archivists moved into the consolidation phase and prepared to distribute the materials for physical arrangement. Instead of dividing all the collections at once, the team decided to start small, with each archivist taking on two collections from the subject list. This choice allowed the team to develop best practices, estimate timeframes for processing collections, and discover other challenges that could surface during the intellectual arrangement and creation of finding aids in ArchivesSpace. Each new subject-based collection required the archivists to pull materials from various locations in the stacks, organize the materials both physically and intellectually into the new subject-based collection, and transfer pertinent finding aid data from the previous small collections into the new, subject-based finding aid. The archival access coordinator then edited, approved, and published the new finding aids. Once physical consolidation was complete, digital consolidation was the final step in the project.

Finding Aids

Throughout the years in Special Collections, processing archivists most often described collections to the folder level and used Describing Archives: A Content Standard (DACS) to govern descriptive format, while Digital Services included item-level metadata description using a defined metadata schema (MODS, Dublin Core, etc.) within its digital collections. As part of the physical consolidation process, archivists planned to create one single collection number for each subject-based collection, thus eliminating all the previous identification numbers for the smaller individual collections under its umbrella and breaking the fragile link between the physical and digital that had existed through the collection number and the EAD Identifier in ArchivesSpace. The problem was that the EAD Identifier field was only available at the collection level, not at the file level within ArchivesSpace (see Figure 3). The compromise was to put a general note at the file level in the finding aid and rename its Local Identifier, a name already in use by Digital Services to describe its digital content (see Figure 4).

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

File Storage and Digital Preservation

Now that the new consolidated collection finding aid maintained the Local Identifiers for previous individual collections, Digital Services needed to determine how the introduction of a new collection number for a consolidated collection affected its file storage of images and metadata on the server and the digital preservation of previous individual digital collections, each with its own collection-level identifier. To understand what changing the collection number would do, it is important to examine the current file storage and digital preservation practices already in place. As part of its digital preservation practices, the University of Alabama Special Collections engages in bit preservation by managing multiple copies of the data, abiding by fixity, and ensuring the data are secure from corruption, damage, or deletion.²⁸ After the creation of a new digital collection, Digital Services stores images and metadata on a network drive before moving the digital content onto a Linux server. Once uploaded to the server, Digital Services generates MD5 checksums for the digital content as a means of ensuring the data are the same and monitors the fixity of the content.²⁹ Digital Services ingests the digital content into the Alabama Digital Preservation Network (ADPN), the statewide LOCKSS network, which replicates the data to seven storage sites.³⁰ With the physical consolidation uniting items thematically under one collection number, the new collection number did not match the old collection number of the existing digitized materials on the server and within ADPN. The choice was either to rename the digital files and metadata and create a duplicate set of files with the new collection number within the server and ADPN, or to leave the digital files and metadata stored under the original collection number. The first option violates best practices of digital preservation, especially fixity. Renaming the files and consolidating the preserved metadata files into one set of metadata files would be time-intensive and prone to human error. Copying and transferring files makes them more susceptible to damage and possible loss of bit-level integrity. This option would also waste storage space with the creation of duplicate files. The second option was to leave digital files and metadata under their original collection number, which seemed the least disruptive option for file storage and digital preservation. With the choice of the second option, Digital Services had to determine how to tie the new collection number back to the original collection number with a “breadcrumb” to navigate within its file storage and digital preservation systems.

After deciding that each individual digitized Wade Hall collection would keep its existing collection-level identifier within its file storage and digital preservation systems, Digital Services brainstormed how to incorporate a new collection number into its metadata template while retaining the original collection's relevant metadata. A new metadata field, the Parent Collection Identifier, will house the new collection number and will serve as a unifying metadata field for previously digitized collections and any materials from the consolidated collection that are digitized in the future. As materials were physically consolidated into new subject-based collections, the archivists created a spreadsheet (see Figure 5) to update the following metadata fields: Collection Name, Item Location (Box and Folder), Parent Collection Identifier, and Finding Aid URL. The Local Identifier represents the unique identifier of a collection at the item level, which Digital Services uses to locate all digital content filed under its previous collection number.

View Figure

• Download as Powerpoint
• Download Figure

Digital Services will use the existing item-level metadata from each individual collection (i.e., Bill H. Axelby letter, u0003_0002039) and add in the updated metadata fields from the archivists' spreadsheet for the new collection (i.e., Wade Hall collection of World War II materials, u0003_0004253) (see Figure 6).

View Figure

• Download as Powerpoint
• Download Figure

Digital Services will archive the updated metadata under its former collection number, with the Parent Collection Identifier field preserving a crosswalk from the legacy metadata in the old collection to the new collection.

Metadata Harvesting and Multi-institutional Aggregators

Several collaborative websites, including Alabama Mosaic, a statewide digital collection consortium EBSCO's Electronic Discovery Service (EDS); WorldCat; and Digital Public Library of America had harvested Acumen using Open Archives Initiative Protocol for Metadata Harvesting (OAI-PMH). The harvest included numerous Wade Hall digitized items. The Wade Hall Consolidation Project and the migration to CONTENTdm overlapped in terms of timing, and it was not advantageous to switch metadata harvesting to CONTENTdm until an overwhelming majority of the content had been migrated. Luckily, the University of Alabama Libraries have maintained persistent URLs to their digital content. After a digital collection has been migrated, Digital Services updates the PURL database to point to CONTENTdm, which provides a seamless experience for any user who clicks on a PURL anywhere on the Web. The most problematic website was Civil War in the American South (http://american-south.galileo.usg.edu), which the Association of Southeastern Research Libraries (ASERL) launched in 2011 to commemorate the sesquicentennial of the Civil War. The University of Georgia's Digital Library of Georgia hosts that website and, in turn, shared that metadata again when its institution became part of DPLA. These harvested websites contained only a specific subset of digital collections dealing with the Civil War, including collections that were part of the Wade Hall Consolidation Project. Digital Services weighed the amount of effort and time it would take to fix the metadata harvesting (with little or no return for both Digital Services and the personnel of these two websites) and determined it was not worth it (see Figure 7).

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

Although seemingly counter to best practices, this conscious choice not to fix metadata harvesting may be another option for institutions that use persistent URLs and are migrating to a new DAM.