Power Through Disparate Datasets with Discovery Interfaces
Jinfo Blog

By Jonathan Jacobsen

Abstract

Jonathan Jacobsen explains the lure of resource discovery interfaces for cultural collections researchers and others. He explores why functions such as faceted browsing and suggestions of related resources, together with the ability to search multiple datasets from one interface, make discovery interfaces one of the hottest trends in web-based search.

Item

Discovery interfaces are one of the hottest trends in providing web-based search access to cultural collections, including library catalogues, journal articles, museum artefacts, photographs, artworks, oral history recordings, historical manuscripts and more. Pioneered by leading websites such as Amazon and eBay, discovery interfaces have been adopted by many public and academic libraries over the past three to four years, though less so by smaller libraries and cultural organisations.

With a discovery interface, users begin a search with a few keywords and the discovery layer guides them to relevant materials. This is accomplished through features such as:

• Spelling corrections and "Did you mean?" suggestions of alternate terms
• Sophisticated algorithms to deliver highly relevant search results
• Faceted browsing, in which the user narrows their search by selecting topics, authors, publication date ranges, material types, and other key information
• Display of related resources (i.e. “If you liked this item, you might also like these”).

These features are intuitive for users and also quite familiar as they are common to so many web applications today.

Ingredients for a "Best of Breed" Public-Facing Search Engine

Discovery interfaces typically also provide options to allow users to act on the records they retrieve, such as:

• Saving searches or subscribing to an RSS feed to keep up-to-date with new additions to the database on certain topics
• Saving, bookmarking, emailing and sharing records
• Adding tags and comments to records
• Exporting records to bibliographic management applications such as EndNote and RefWorks
• Simultaneously fetching and incorporating external content, such as author biographies from Wikipedia, reader reviews from Amazon and book covers from Google
• Indicating loan and availability status and allowing holds to be placed
• Sending a citation via SMS text message to a mobile device.

As a technology, a discovery layer always sits on top of existing systems. It’s not intended to replace an integrated library system or cataloguing application, but rather to be a "best of breed" public-facing search engine.

Discovery interfaces are often not limited to searching just one dataset. A single interface may index multiple library catalogues, digital repositories, websites or other external content. With this approach, information is brought out of silos and users need only use a single interface to access disparate resources. Some commercial systems are available pre-loaded with content, such as journal articles from databases and journals to which you already subscribe. To these you may add your own local content, such as your catalogue. Other systems are empty, ready to index your library catalogue records or other content.

Discovery systems for libraries are available from numerous commercial vendors, as well as a few open source communities. Selecting and implementing a discovery interface is not necessarily challenging, but requires a thorough process to select the most appropriate system and the most appropriate data to include in it. A discovery interface may include data from multiple sources, but the data may need to be manipulated and the interface configured to provide the best experience for users.

FreePint subscribers can read more about discovery interfaces. Log in to view Upgrading to a Resource Discovery Interface.

« Blog