Marshall Breeding, long-time attendee and contributor to both the IL and CIL conferences, gave us an update on discovery services.
Discovery services are among the most important products that libraries can invest in. Most traditional OPACs include books, journals, and media at the title level. Without a discovery service, articles, book chapters, ad digital objects are not available. The initial discovery products focused on interface improvements. Present-day discovery service interfaces include a single search box, query tools, relevance-ranked results, faceted navigation, enhanced visual displays, and recommendation services.
The current phase is focused on pre-populated indexes that aim to deliver Web-scale discovery. It is very difficult to develop a system that mimics Google. Discovery services depend on a consolidated index of sources available to a library. As a website replacement, discovery services provide a portal environment including customized content management services that can fulfill requests for typical offerings on library websites, and full integration between an existing website and resource discovering.
Scalability to billions of items is a challenge. How do we handle things on a Google-like scale? (Nobody has ever gotten a “Google is full right now. Please hold” response to a search query!) Libraries must also be able to handle user-generated content. Another major problem is comprehensiveness. Technically, we can index everything a library subscribes to, but there are political reasons we cannot do so. Many publishers, particularly A&I service providers and aggregators, object to their content being hosted in a discovery services because of the effort they have put into compiling their databases. And many competing services will not give their data to other services.
The indexes in discovery services are not just citations; full text indexing enhances access. It is therefore important to understand in-depth indexing. But the relevancy of searching such a mass of data can be a problem. Here are some challenges for relevancy.
Discovery should be a more social experience, but it takes lots of content to be effective. BiblioCommons and ChiliFresh are two providers working on this problem. E-book integration is important for public libraries, and some libraries are loading MARC records for e-books into their local catalogs, so that discovery and lending operations (i.e. downloading to a user’s reader) can be performed fully within the library’s discovery service. Each e-book service provider’s API is inconsistent with others, posing large challenges for system developers.
Currently, discovery services are a “must have” for academic libraries, but public libraries are moving back to a discovery interface associated with their ILS.
Breeding has authored a recent report on discovery services in Library Techology Reports (LTR), which contains extensive data on library’s opinions of discovery services.