Implementing Book Search and Catalog Search

Scott Frey, reference library at Western State University College of Law, says he has both law and library background but claimed not to be a trained IT specialist. He started by asking why a librarian might want to develop a search engine for econtent. The obvious answer—to collect book resources from various sites onto one site. Other reasons might be to tailor a search tool for researchers that the librarian understands and to better understand both search engines and web pages, which could lead to better web research skills.

He also stressed why librarians can’t just rely on Google. While Google is fast and useful for all kinds of econtent, it makes mistakes with the OCR text and the metadata, and may not capture or display results the way librarians want. He listed some technology issues that might be involved—insufficient hardware or bandwidth, buggy software, security issues, and more. There are also bibliographic issues (even the definition of ebook) and legal issues (copyright, breach of contract, etc.) that must be considered. Then he discussed some of the component options (software for crawling, indexing, and searching—Nutch and Webglimpse are tools that do all three). His presentation slides also included a list of Web sources for full-text ebooks (available in the conference presentation volume and will be posted to the conference site).

Maria Armitage and Amy Barnes talked about their experiences of implementing new catalog search at Columbus Metropolitan Library. The library wanted a solution that could sit on top of their existing home grown ILS solution—and chose AquaBrowser. Search results are now offered with relevancy ranking and boosting of most popular items. It also provides faceted search navigation. RSS feeds can be customer customized. Fuzzy searches help users get to the desired terms—a “did you mean?”. Rollout took only about 3 months.

Their rollout to the public was not completely smooth, so they presented a list of lessons they learned.

  • Don’t just use vendor defaults
  • When things go wrong, fix it fast!
  • Know your numbers – session and query data, etc.
  • With faceted interface, little anomalies become magnified – "MARC magnified"
  • Catalog users are not necessarily the same as web users – expect some surprises
  • Buy-in is so critical, for both staff and customers (communicate, provide early looks, a transition, knowledgeable staff, etc.—and emphasize the WOW factor)

Future plans include adding federated searching, database integration, social networking tools, events, etc., as well as a kids catalog. 

Background note: Medialab is the creator of the library search and discovery platform AquaBrowser (www.aquabrowser.com). AquaBrowser is an ILS vendor-neutral, visual faceted search that connects to any number and type of data sources.  Medialab Solutions BV is dba AquaBrowser North America in the U.S. and is a business unit of R. R. Bowker (www.bowker.com).

Paula J. Hane
News Bureau Chief, Information Today, Inc.
 

Tags: ,

Comments are closed.