LTS Newsletter: July 2025

Library Technology Services, Harvard Library

Welcome to Library Technology Service’s occasional newsletter for anyone interested in Harvard Library IT.

Coming Soon: New Viewer

A newer version of Mirador, the image viewer implemented by Harvard Library, will soon be rolling into production, bringing a host of improvements to accessibility and user interactions. This new release introduces features such as enhanced keyboard navigation, better screen reader compatibility, and flexible display options to support users with varied needs. The refreshed interface also offers more intuitive controls for studying high-resolution images.

Stay tuned—further details about the new Viewer’s rollout to discovery platforms will be shared soon. Because this implementation involves decoupling components from our legacy delivery systems, it will be handled incrementally and accompanied with advance notice to stakeholders and users.

OpenAthens final phases in August

The final two phases of the OpenAthens migration project are nearing their completion this August! To refresh your memory, OpenAthens is an authentication system that provides single-sign-on access for subscribed electronic resources (e.g. journal articles and e-books). It is a replacement for EZProxy and offers more flexibility in how users are authenticated. Currently there are approximately 160 vendor platforms already using OpenAthens. Users who access library resources through those platforms are directed through OpenAthens, rather than EZProxy. They continue to login via HarvardKey for access.

Phase 3 of implementation is scheduled to go live August 8-9, 2025. This phase includes the majority of resources that remain on EZProxy (approximately 550 vendor platforms), with a few exceptions reserved for the final phase. A full list of resources is available on the OpenAthens page on the LTS wiki. Phase 3 resources will continue to use IP-based authentication and will be proxied through OpenAthens. For these resources, EZProxy links will continue to work for now, though library staff should be updating links to use OpenAthens. Guidance on updating links can be found on the wiki page for Updating links to OpenAthens.

The Final Phase of OpenAthens implementation is planned for the week of August 11, 2025. These resources must be implemented last because the vendor requires a hard cutover from EZProxy to OpenAthens. They cannot support multiple authentication methods at the same time. This list includes:

Ebook Central
Pivot
Gale
The Economist

For these resources EZProxy links are not expected to continue working, so please update those links accordingly, either to point to a HOLLIS permalink or using the guide linked above.

Once the final phase of migration is complete, there may be some platforms that remain on EZProxy for a time, often due to communication issues with the vendor or technical issues that haven’t resolved yet, and we will continue to work with those vendors. However, we do expect the majority of the 600-700 platforms currently used across Harvard Libraries to have migrated from EZProxy to OpenAthens. Please look to the OpenAthens page on the LTS wiki for updates and more information.

AI-Augmented Software Engineering Pilot

The Library Technology Services (LTS) Data & Infrastructure and Discovery teams have launched a pilot project to explore how AI can support and enhance software development practices. This is an exciting time for the profession, as large language models (LLMs) are becoming increasingly integrated into the tools developers use every day, opening up new opportunities to boost productivity and free up time for strategic work and innovation.

Just as tools like Google Sheets and Microsoft Excel support spreadsheet work, software engineers use specialized tools to write code. One such tool, Visual Studio Code (VS Code), is widely used by LTS developers. A plugin called Cline enables VS Code to connect with large language models, bringing AI directly into the coding process.

In collaboration with the Apigee team, LTS engineers are using this AI-augmented environment to evaluate the benefits of different LLMs, focusing on time savings and the quality of the code produced. Apigee helps route requests to the AI models and provides detailed reporting on usage and cost.

The results of this pilot, measuring cost, efficiency, and effectiveness, will help inform whether this AI-assisted approach should be expanded to more of the LTS team.

Archives Space Data Corrections

Digital Object Cleanup

Approximately 9,000 Digital Objects in ArchivesSpace had incomplete XLink attributes. When exported via EAD XML, the finding aid would be invalid, thus blocking the collection from appearing in systems like ArcLight and LibraryCloud. Due to the scope of the issue, a system-wide cleanup was warranted. LTS identified these Digital Objects and applied the correct XLink attributes.

Ref IDs

24,000 Archival Objects in ArchivesSpace had malformed Ref IDs. These are unique identifiers for Archival Objects and relied on across Harvard Library systems. Improperly formed Ref IDs can cause issues with id.lib.harvard.edu links, ArcLight publishing, and LibraryCloud transformation. Ref IDs are not directly editable within ArchivesSpace, leading to their buildup over the lifespan of the application. LTS identified records with bad Ref IDs and regenerated them on July 29th. A second job will be run to regenerate Ref IDs that inherited irregular EAD IDs.

LTS was happy to partner with the ArchivesSpace Service Owner, Alex Duryee, to identify and process these data corrections to ensure that ArchivesSpace and adjacent systems can rely on accurate data.