Digital Access

As part of the Content, Context & Capacity grant's experimental approach to large-scale digitization it was decided that -- at least initially -- the online finding aids would serve as the main context and discovery platform for digitized content and that additional metadata (other than what could be automated out of the finding aids via scripts) would not be manually crafted for digitized items. Therefore, each university library links digital content directly from online finding aids at the container-level component, and all scans related to a given container-level component have the same basic metadata (for example, the collection, series, and container titles).

While each of the partner institutions hosting digital content has its own finding aid and digital content publication platforms, all finding aids and digital content are accessible not only through individual institutions' websites, but also through Search TRLN, which enables searching over combined TRLN collections from within a single software interface. TRLN hosts copies of each institution's finding aids but does not host copies of digital collections. This means that while finding aids from all four consortial partners have a similar look and feel in the Search TRLN interface, links from finding aids to digital content will take users to one of three digital content delivery platforms: that of Duke, NCSU, or NCCU/UNC-Chapel Hill.

View a video tutorial describing how to access digitized content through the finding aids at TRLN institutions.

Methods for publishing and linking digital content from finding aids

Duke Libraries

Duke is encoding <dao> elements at the component level for each component with digital content and delivering finding aids via a custom Django web application. During digitization, each EAD component-level container is assigned an identifier, which is recorded in the finding aid. Image filenames use a concatenation of the container identifier and a sequential number representing the scan sequence. The <dao>'s href attribute is the container id. A python script parses the finding aid <dao> elements and generates a METS file for each container with digitized content. Each METS file includes title and series metadata for the container, a <structMap> that contains a <div> for each scanned page in the container, and paths to the different size JPG derivatives in the <fileSec> for each component. When a user clicks the container-level link produced in the HTML display by the <dao>, an AJAX call is made to retrieve a block of inline HTML for the container, including the thumbnail image derivatives for the first 50 scans in the container, and a hidden placeholder for any pages beyond 50. Mousing over a thumbnail pops up the original thumbnail in a tooltip using the TipTip jQuery plugin. Clicking any thumbnail launches a lightbox-style slideshow using the Colorbox jQuery plugin to navigate the images. Each container with a <dao> is assigned an HTML id attribute to enable direct linking via a URL hash. If the page loads with a URL hash, the 'click' event fires for the <dao> link, thus displaying the thumbnails for the container.

See a live example of Duke's finding aid to digital content linking.

Figure 1. When a finding aid links to digitized content, a message appears at the top of the finding aid

Figure 2. If the contents of a container have been digitized, the words "View digitized items" will appear as a link.

Figure 3. Click on the "View digitized items" link to see the first 50 thumbnail images

NCSU Libraries

NCSU is encoding <dao> elements at the component level for each component with digital content, and delivering digital content through a Ruby on Rails application utilizing the Blacklight plugin for Solr-powered search. The application reuses the MySQL and Solr databases managed by the NCSU Special Collections Asset Management System. This application was an early adopter of HTML5 Microdata and Schema.org for exposing embedded, structured data and improving discoverability on the open web. Other technologies used in digital object display include jQuery and the Djatoka image server. The interface allows for resource discovery through search and faceted browse.

Each container-level component in the EAD finding aid is assigned a semantic identifier based on the folder's intellectual and physical location in a collection. The identifier includes components representing the EAD identifier, series number, container number/s, and an empty resource sequence number. The container-level component identifier is the base identifier for all scans that will be made from the folder. Using this base identifier, brief records are automatically created in the NCSU Special Collections Asset Management System for container-level resources that have been digitized. Brief records include titles harvested from EAD collection and component <unititle> elements, as well as some manually assigned elements determined at a batch level. Once metadata records exist, images are processed into JPEG2000 files, which are loaded onto a djatoka image server. Once a record is associated with images, records are indexed and availaible in a digital collections platform. Links at the component-level EAD container are entered into the collection record in Archivists’ Toolkit (AT). The XML record is exported from AT and moved into a directory on a web server. This directory is re-indexed hourly; after re-indexing the updated live collection guide will include links to the appropriate digital content in the digital collections platform. A prominent link label will appear at the top of the "Contents" tab of the collection guide stating, "Portions of this collection have been digitized and made available online. The entire collection, including materials not available online, may be viewed in the Special Collections reading room in D.H. Hill Library." We also auto-generate an "Online Content" tab in the collection guide, which lists the containers that have digital material online.

See a live example of NCSU's finding aid to digital content linking.

Figure 1. When a finding aid links to digitized content, this appears at the top

Figure 2. If the contents of a container have been digitized, the words "Online content" will appear as a link.

Figure 3. Click on the "Online content" link to be taken to a listing of all digital content associated with the container

UNC-Chapel Hill and North Carolina Central Libraries

UNC-Chapel Hill and North Carolina Central are using CONTENTdm to delivery digital content, but they do not encode <dao> elements in the EAD finding aids. Instead, the XSLT script that transforms EAD finding aids to HTML adds an identifier to each component level that has a container type. A second XSLT script creates the same identifier as part of a spreadsheet containing minimal metadata to upload to CONTENTdm with each digital object. When an HTML finding aid is downloaded by a browser to display to a patron online, a javascript is envoked that crawls both the digital content in CONTENTdm and the HTML finding aid. The javascript dynamically creates hyperlinks at the finding aid component level if digital content with a matching unique identifier is found in CONTENTdm. No extra encoding or modification of finding aids is required in order for links to newly digitized content to appear. This client-side, real-time approach ensures that as new containers in any collection are digitized and loaded into CONTENTdm, they become immediately available to users through the finding aid, without any further intervention from staff. At the same time, this digital content linking method relies heavily on consistent practices in EAD encoding for thousands of finding aids. Without consistent encoding practices for container types, container numbers, and collection IDs, an automated process would not function well.

See a live example of UNC-Chapel Hill's finding aid to digital content linking. This method is also used for NCCU's finding aids.

Figure 1. When a finding aid links to digitized content, this appears at the top

Figure 2. If the contents of a container have been digitized, the container number will appear as a link, followed by the phrase, "(digitized)".

Figure 3. Click on the container link to be taken to a search results page in CONTENTdm listing all digital content associated with the container