Written for Unlimited Priorities and DCLnews Blog.
A major topic at current conferences is “discovery,” and this was certainly true at the recent Internet Librarian (IL) conference in Monterey, CA on October 25-27. So what is discovery and how does it differ from other forms of search?
Information users are trying to access more and more types of information, both in locally stored databases like library catalogs, and external information from commercial systems, and increasingly they want to do it with a single Google-like search box. The problem is that content is often siloed in many different databases, and it comes in a wide variety of formats. Most users therefore become frustrated in finding the information they need because they do not know where the information resides, what the database name is, and how to access it.
The first attempts to help users took the form of “federated search,” in which the user’s search query was presented to several databases in turn, and then the results were aggregated and presented in a single set. The problem with federated search was that the user had to select the databases to be searched, and then had to wait for the sequential searches to be performed and processed, which could result in long response times.
In current discovery systems, a unified index of terms from a wide variety of databases is constructed by the system, and the queries are processed against it. Several discovery systems can aggregate the results and remove duplicate items. This approach has significant advantages over federated search systems:
- The user does not need to know anything about the databases being searched,
- Because only one search is performed, searches are faster, and
- The same interface can be used to search for information from all databases.
A major issue with discovery systems is that considerable effort (and time!) must be expended in installing a system in an organization and customizing it to access only those systems to which access is available.
Activity in discovery systems in currently intense, and several competitors are vying for market share. Examples are Summon from ProQuest, EBSCO Discovery System, WorldCat Local from OCLC, and Primo from Ex Libris. One might ask if discovery systems are better than Google Scholar. One panel at IL looked at this question, and the answer is that they appear to be, but further investigation is needed. And because these systems are newly developed, problems with installation and the relevance of results have not all been solved yet. But it is clear that discovery is an exciting new advance in searching, and we can expect to see new advances coming rapidly in the near future.