Shared Discovery Services Working Group
The Shared Discovery Services Working Group is responsible for the design and implementation of a shared discovery layer to replace Endeca. The scope of this work includes the replacement of Search TRLN and all local member library discovery layers built on the Endeca platform. Working group members have the responsibility to communicate and consult with staff at all of the libraries at their institution regarding the group’s efforts.
The working group embraces an “Agile Development mindset” and follows the Scrum framework, with an Advisory Team, Product Owners at each institution, a Scrum team, and a Scrum Master. More details about the application of Scrum to this working group’s activities are available in the Proposal to Apply Scrum Framework for TRLN Discovery.
Under development. Contact Info@trln.org with any questions.
Working Group Documents and Resources (May include work plans, quarterly work reports, meeting minutes and other resources developed by this Working Group)
More About the Project to Build a Shared Discovery Layer
“TRLN Discovery” is a collaborative software development project to implement a modern and open-source discovery service which allows the user to conveniently find and access materials from all TRLN member libraries within a single search environment. It is unique in uniting the catalogues of multiple independent institutions who are not part of the same system. Learn more about the project below. Questions or requests for more detail may be directed to firstname.lastname@example.org.
TRLN Discovery will maximize the value of libraries’ collections by developing a conveniently managed and customizable discovery service that enables users to effectively and efficiently locate and access research materials from the Duke, NCCU, NCSU, and UNC libraries.
In late March 2016, TRLN announced its decision to replace the existing Endeca-powered library catalogs and Search TRLN with a new discovery service aligned more closely with TRLN’s current context. This decision was reached in part due to Oracle’s cessation of development of the version of Endeca used for Search TRLN. While Oracle encouraged TRLN to implement the new version of Endeca set to launch in Spring 2016, they clarified that the effort involved in installing and implementing a new version of Endeca would equal or surpass the effort invested in developing Search TRLN initially.
Through conversations with various TRLN councils and committees, TRLN decided to pursue the implementation of a new discovery platform, with Blacklight and VuFind identified as high-level candidates. Two groups were formed to lead the development of this project — an Advisory Team with membership representing the Technology Council, Services Council, and Collections Council, to oversee the project timeline and goals and liaise with campus stakeholders on major decisions, and a TRLN Discovery Steering Committee responsible for project management and technical implementation. Three smaller Implementation Teams — User Experience and Services, Data Extract and Ingest, and User Interface Development — reported to the Steering Committee.
After consulting academic libraries and library consortia, conducting background research, and completing group and individual evaluations, the TRLN Discovery Steering Committee recommended that the Blacklight open source discovery system be used as the foundation for the new TRLN discovery service. The committee felt that both Blacklight and VuFind would fulfill the requirements of this project but noted that Blacklight’s modularity and extensibility would more readily facilitate collaborative development while also allowing local variation when necessary.
In May 2017, at the recommendation of the Advisory Team, TRLN implemented an Agile approach, specifically Scrum, for the implementation phase of the project. The teams and committees that had been involved with the project up to that point were reconfigured or dissolved as necessary in order to align with Scrum principles. TRLN financial and staffing resources went into the purchase of Jira and Confluence for project management.
The product is hosted by Amazon Web Services (AWS) in the cloud; this is TRLN’s first experience with cloud hosting.
Base implementation is planned for Summer 2018. The launch date for institutional customization is January 2019.
October 2018 Update:
- The implementation team identified a way to manage shared record sets (one university loads records on behalf of all) and is working toward replicating the current shared record sets for the January production launch.
- The team added an ability to sort values in hierarchical facets by count. This will allow the location facet to display libraries with more items at the top of the list.
- A separate Argon sub-account and AWS architecture are being developed for NCCU. The AWS architecture will be evaluated by an AWS architect on October 31.
- Local teams have increased their efforts to prepare their local Blacklight skins for the production launch. Those efforts have resulted in several design improvements for the user interface.
- Work is underway to implement accessibility improvements before launch informed by accessibility testing performed at UNC.
- Additional usability testing was conducted at Duke to evaluate how users interacted with catalog facets and provide recommendations for future changes.
September 2018 Update:
- Per the Advisory Team’s recommendation, the implementation team met with a consultant from Innovent Solutions on September 6 and 7 to review the project’s Solr implementation. Results of the 2-day meeting included a strategy for developing relevance tests.
- This month the development team finished the autocomplete feature and reduced the number of facets on the landing page.
- Local instance testing continues at Duke, NCCU, NCSU, and UNC.
August 2018 Update:
- The TRLN Discovery development team delivered a presentation about the project status and outlook at the TRLN Annual Meeting.
- Local instances of the new discovery system are now available for testing at Duke, NCSU and UNC. Library personnel are encouraged to test the system and provide feedback. NCCU will have a local instance available for testing in September.
- The TRLN Discovery implementation team will meet with a consultant from Innovent Solutions on September 6 and 7 to review the project’s Solr implementation.
- Major features being worked on this month include request (local document delivery and interlibrary lending) and search auto-suggest.
July 2018 Update:
- The TRLN Discovery development team started working on a presentation for the TRLN Annual Meeting where they will offer a project update, answer audience questions, and gather reactions to the internal alpha release scheduled for the first week of August.
- The product owners have reviewed and prioritized the features planned for the public launch in January 2019. They have identified the highest priorities so that they can begin making immediate progress after the internal alpha release.
- The Advisory Team agreed to a recommendation from the development team to engage a consultant to review the project’s Solr implementation and to help the team make sure the index is efficient, reliable, and ready for the public launch in January 2019.
- The development team participated (along with other information technology staff from TRLN libraries) in an AWS Well-Architected Framework Session lead by an AWS Solutions Architect. The session offered advice and best practices for hosting systems on the AWS cloud.
June 2018 Update:
- The TRLN Discovery team has completed initial rounds of user testing and has begun to incorporate feedback into the user interface.
- UNC has completed a full record load from their ILS into TRLN Discovery’s solr index. All four universities will complete full record loads for the August 1 internal release.
- The development team has scheduled an AWS Well-Architected session with an AWS solution architect to assist in a review of the system infrastructure.
- The product owners have scheduled a meeting for July 20 during which they will prioritize the work for the upcoming public release in January 2019.
May 2018 Update:
- On May 22, Genia Kazymova attended an all-day AWS Administrator training in Washington DC.
- The team has continued to refine the user interface with significant contributions coming from Michael Daul at Duke. The team remains optimistic about debuting the service in August.
- Genia Kazymova has been working with Yan Wang at NCCU to ingest NCCU holdings into TRLN Discovery. The work has been an excellent introduction for Genia to the complex world of MARC bibliographic data.
- The team has continued to make improvements to the infrastructure hosted on AWS in preparation for storing a much larger set of records for the August launch.
April 2018 Update:
- On April 16, Genia Kazymova joined TRLN in the position of Applications Developer. She is spending her first few weeks getting up to speed on the TRLN Discovery project and getting to know the TRLN Discovery team.
- The TRLN Discovery development team added an extra hour to their regular meeting schedule to make more time for backlog refinement and the discussion of major issues that require both developer and product owner input. This extra time helps ensure that they will meet their goals for an August 1 internal launch date.
- The team has continued to refine the user interface in preparation for the August launch.
March 2018 Update:
- The TRLN Discovery team took the first step in scaling up the infrastructure on AWS, since the project is now at a stage where the number of records ingested into the system required a larger machine. This is part of the effort to scale up to a much larger number of records to be ingested and indexed for the Alpha release in August 2018.
- The team has also implemented SolrCloud to support fault tolerance, load balancing, and redundancy within the discovery infrastructure.
- The team continues to make great progress toward the Alpha release set for August 2018. The public interface is also receiving special design treatment by Michael Daul and Sean Aery from Duke University Libraries. The team is excited about the design improvements that are being made and looks forward to showing off the system this summer.
February 2018 Update:
- The TRLN Discovery team set a firm date of August 1, 2018 for the launch of an Alpha version of TRLN Discovery. The Alpha release is intended for staff at TRLN libraries only. The August 1st date is firm and features that are not complete by that date will be moved to the Phase 1 release rather than delaying Alpha.
- The TRLN Discovery team met for an all day meeting on February 26th where they completed the refinement and prioritization of all features currently set for the Alpha release.
- The TRLN Discovery team participated in interviews with three candidates for the vacant TRLN Applications Analyst position during the last week of February. Since the position was posted on January 22, we have been working to identify and recruit a qualified candidate as quickly as possible.
January 2018 Update:
- The TRLN Shared Discovery Advisory Team met virtually on January 18 to review project status and discuss challenges. Jill Sexton (NCSU) joined the Advisory Team.
- Project team has moved forward, despite fall 2017 departures of the TRLN Applications Analyst/Scrum Master and UNC’s Applications Analyst.
- Both Duke and NCSU are running local Blacklight instances to facilitate testing and development.
- The project team will hold a 1-day summit on February 23 to refine and prioritize all remaining features that will be included in the first release to library staff.
- The project team expects to launch the first Alpha release of the software for library faculty and staff testing on August 1, 2018.
- The team remains committed to the work and mindful of resourcing and timeline.
- UNC is in the process of hiring another developer. TRLN’s Applications Developer position has been posted and applications are currently under review.
November 2017 Update:
- In August 2017 we had our first release, which was a skeletal version of Blacklight on top of a Solr instance hosted by AWS. This release was for Product Owners only.
- In January 2018 we hope to have our second release. This will be a version of Blacklight ready for library staff to review.
- We currently have 300,000 records in Solr. We hope to scale up to 2 million records for the second release.
- Some of the second release features already achieved include item availability, record rollup, integration with Syndetics TOC and summary data.
- Shashi Penumarthy, TRLN Applications Analyst, departed TRLN in July. Ben Heet, TRLN Program Officer, has been serving as interim scrum master.
- TRLN has completed a job description for an EHRA Applications Analyst and hopes to post the position as soon as HR processes allow.
- TRLN is working with MCNC to extend the current contract for hosting the Endeca application through January 2019. We plan to use one year contract extensions until the new discovery service is launched.
- The Development Team, including the Product Owners, remain optimistic about the project and the work being done.
- The projected dates for project completion are subject to change. Shashi Penumarthy’s departure from TRLN and the subsequent vacancy may require an extension of the project timeline.