8 steps for building a better digital Library experience

The Library has released a preview of our new catalogue which is now available for public testing and feedback. As part of this ongoing development process, we want to share our journey with you, including challenges, decisions and learnings. As well as sharing more information about why we needed a new catalogue experience, here are 8 crucial steps that we’ve taken to make this project possible and (hopefully) successful.  

1. Reduce technical terminology; harmonise reader-focused language 

The first step was to minimise confusion caused by library terminology.  

Early in the project, we conducted interviews with users of our existing catalogue both onsite and remotely, observing how they used the system and interpreted information about the Library’s collections. Over 150 readers have been engaged through this journey, from specialist researchers to casual Library visitors. 

Terminology used to describe collection formats was a particular pain point for readers. Our current catalogues together have over 30 different format types. 'Ephemera' and 'Manuscript Maps' came out as the most confusing.  And 'Print' without context was vague – did this mean printed text or an image print? We have now reduced the format categories to just 11, using simpler, less 'library-ish' language. 

With this in mind, we are now looking for similar ways to make our copyright and access terminologies consistent and simple for the reader. 

2. Going headless 

We had quite a bit of fun explaining the title of this step to our executives and staff!  

A quick intro: 'going headless' comes from the trend in content management systems where the presentation front-end ('the head', or what the user sees) is separated from the content back-end ('the body', or what administrators see). Rather than these two functions being tightly coupled, going headless gives you the reliability of working with fit-for-purpose back-end systems but the flexibility to custom design the front-end to suit the needs of your users. 

We took this approach to retain the core functionalities of our current systems that do their intended job well. Our collection management staff will continue to use these 'systems of record' which hold authoritative collection information according to library, archival and preservation standards.   

Library staff are now taking the lead in designing these digital interfaces to best suit the needs of our readers, moving away from the commoditised and generic interfaces found in many library catalogue systems. 

3. One data model to rule them all 

The next step was to bring the data from multiple systems together in a way that enables us to normalise and standardise the data, without the need to change data in the source system. 

This unified data structure enables us to present all collection information from many systems, such as bibliographic records, archival hierarchies, and digital images, on one page seamlessly. We hope to continue to enrich this collection data model in the future by connecting related stories from our website, upcoming events, and even items available from the Library Shop. 

4. Decoupling preservation and access 

Back in 2015, we implemented a system to meet our digital preservation, digital asset management (DAM) and digital delivery needs. In the short span of four years, the digital preservation domain has become more sophisticated: the domain knowledge, best practices, vendors’ systems, and staff capabilities. 

Our digital preservation processes at the Library are meticulous and robust, but while the system we had chosen to manage this meets our preservation requirements, it has continually struggled to meet reader and staff needs for accessing high quality digital files easily and quickly.  

The decision to systematically decouple our digital preservation and access functions was a difficult but necessary step to achieve our goal. Difficult in that migration and re-ingestion of millions of files was required, but necessary in that we can now enable faster and more seamless access for our readers to many of the treasures that the Library holds. 

5. Using the cloud to improve reliability and speed 

With the headless approach, a unified data model, and a new DAM, the next step we took was to re-engineer delivery using modern cloud technologies to enable speed, reliability, and scale.  

Our work is now hosted on a cloud compute environment, which in conjunction with our new API (Application Program Interface) management platform, enables faster performance. This has significantly improved the delivery and performance of webpages, data and images for readers. We are seeing improved load times for images - up to one hundred times faster than the current system.  

Researchers, students and partner institutions are increasingly asking us for large collection datasets and we will now be more able to fulfil these requests quicker and more easily with our new API management platform. 

6. Make collection hierarchies easier to browse 

A catalogue hierarchy is important to communicate provenance of an item, especially for large original materials collections. Structured like a family tree, it describes relationships between sub-collections, series, groups, and items, and is invaluable for researchers wanting to understand the context of collections. Many of our collections are four or five levels deep, with some levels having over 200 items.  

The hierarchy display in our current catalogue struggles with large and complex collection hierarchies. The hierarchies would load slowly and only a few levels at a time. Readers would easily lose their place trying to navigate through the tree while the page was loading. Sometimes in the middle of trying to load new levels of the tree, the hierarchies would crash all together and you’d have to start from the top again. 

In our new catalogue, the Library’s DX Lab took on the challenge to deliver an elegant and simple solution to delivering the important collection hierarchies. You can test the first iteration of their solution on the May Gibbs or Sam Hood collections.

7. Use computer-generated tags to describe images 

The Library has more than 1.5 million photographs and many are not individually described. For example, it is common to find an album of photographs where there is a brief description of the album as a whole, but no description or captions for the individual images. With such a large collection it is all too easy for individual photos to get buried and remain undiscovered.  

Around three years ago the Library started experimenting with the use of artificial intelligence and machine learning to generate automatic tags for our photographic collections. After a lot of testing we have now refined a process, known as TIGER (more on this in a later post), to tag our image collections as part of our ingestion process using cloud technologies. For the first time, readers can use keywords such as 'horse', 'badges', and 'palm trees' to find images from our collections. This technology even works to an extent for drawings, objects and paintings. This is an exciting new way to discover treasures within the Library’s archive and a key feature that powers our new digital collections platform. 

8. Putting it all together and making it beautiful 

Finally, the last step (for now) has been to employ UX (user experience) design principles to deliver new search and digital collection interfaces that are simple and useful to our readers.   

Good, sustainable design means there is a pattern to our approach: components, elements and templates are re-usable, follow design practices and a consistent design language.  We have created building blocks in our stylesheets – elements like colour palette, page grids, icons, and typography, that can be re-used over and over. 

An important part of our design process has been a focus on making our new catalogue work equally well for all people across as many devices, operating systems and browsers as possible. Accessible design has been a key focus and as part of our ongoing work we have conducted testing with Library users who access our services in different ways and who may also use different assistive technologies, such as screen readers. 

We believe that these steps will enable us to provide more meaningful research experiences for our readers and to ultimately bring our collections out from the stacks and into the lives of even more people. This journey for the Library will continue to develop, but we're sure it will support our readers to search and browse the Library's collections more easily, intuitively, and enjoyably. We hope you agree.