American Libraries |
||
Site NavigationLeft Sidebar ItemsOnline Features
Follow American Libraries news stories, videos, and blog posts on Twitter.
|
||
Much Ado About Metasearch
Head of Systems, North Carolina State University Libraries, Raleigh. Column for June/July 2004 The single search box—that simple gateway to a world of knowledge—has been the pipe dream of libraries since the advent of online library services. As with most things, we must be careful what we wish for. But as discussion lists, professional publications, and library conferences debate the utility, shortcomings, and near ubiquity of metasearching, others have realized that the horse got out of the barn a long time ago. One group of professionals, the Metasearch Initiative, is trying to saddle that horse before it gallops away. A new problem, an old solutionIt’s worth it to review some background first. Metasearch, single search, federated search, parallel search, broadcast search—despite subtle differences, these all describe the same thing: the ability to search a heterogeneous set of resources with a single interface. Most of those disparate resources serve their data in many different ways (or not at all); nevertheless, several vendors and libraries are already creating interfaces to broadcast searches to multiple database targets. Patrons have become accustomed to promiscuous web interfaces that will search anything; the challenge for the library profession is to create ways to search well in all the right places. In days of old and new, this was done primarily with Z39.50. This older standard, however, has three fatal flaws. First, it was not really meant for article-level citations. Second, Z39.50 for citation databases was primarily built to facilitate bibliographic citation software (for example, Procite and Endnote) connections for a few targets from only a few users. Finally, Z39.50 remains a high barrier of entry for the vast majority of database vendors—and in particular the primary publishers—who offer no search option other than their proprietary interface. Given Z39.50’s shortcomings, a burgeoning industry has emerged in federated searching. With no official standard to go on besides Z39.50, several vendors have created a cottage industry around creating connectors to various database targets. Also known as “HTML parsing” or “screen-scraping,” the HTTP protocol involves determining where desired data resides on an HTML page so that the metasearch client can display information to patrons in a single interface. This is the kind of work that would be too costly for libraries to conduct on their own. But even vendors recognize the tedious, repetitive nature of the work and prefer a new approach. In search of a standardAs a result of increasing metasearch activity, several vendors also began to notice spikes in search activity. For those without a standard server application like Z39.50, these searches appeared like any other, even though a federated-search agent was performing the search. At the ALA Midwinter Meeting in 2003, three vendors—EBSCO, Gale, and ProQuest—called a meeting to discuss the problem. The National Information Standards Organization (NISO) volunteered to take a leadership role in determining a solution. Two workshop meetings followed. “The Denver Meeting” in May 2003 uncovered the breadth of the problem, from user agent identification, authentication, and authorization to collection discovery and citation metadata. A meeting in Washington, D.C., followed in the fall. Held in conjunction with the completion of the OpenURL standard, this meeting opened the issue to an even broader audience of ILS vendors, middleware providers, librarians, and publishers. As a result, NISO launched the Metasearch Initiative, which will identify, develop, and frame the standards and other common understandings that are needed to enable an efficient and robust information environment. The goal of NISO’s Metasearch Initiative is to enable:
The work of the committee is overseen by chairs Jenny Walker of Ex Libris and yours truly. Three subcommittees have been designated to investigate access management (chaired by Mike Teets of OCLC), collection description (chaired by Juha Hakala of the National Library of Finland), and search and retrieval (chaired by Sara Randall of Endeavor and Matt Goldner of Fretwell-Downing). With a mix of librarians, publishers, and library-software providers, the initiative has drawn the participation of over 60 individuals from five countries. Even large publishers who have already developed proprietary XML gateways are involved. “Elsevier is in the unique position of being both a content provider for libraries using metasearch as well as a supplier of metasearch capabilities,” commented Marc Krellenstein, vice president of search and discovery technology. “We believe that a standard which enriches the user experience through improved metadata and search capabilities while reducing the impact on content providers’ systems is a benefit for all parties.” Great things start smallThe first face-to-face meeting took place in North Carolina’s Research Triangle Park in April and drew nearly two-thirds of the committee members. The success of the meeting is not only measured in the high level of participation and work accomplished, but also in the cooperative effort of disparate parties to solve the issues of metasearching for everyone—libraries, vendors, and publishers alike. Standards are a major tool for library-vendor-publisher cooperation. The rich makeup of NISO’s Metasearch Initiative is just one shining example of a combined effort. While shopping for metasearch solutions or new databases, it might be wise to inquire if the vendor or publisher in question is tracking this important NISO initiative. While it has been said that librarians will make standards while vendors make reality, the Metasearch Initiative is now working toward both at the same time. Contracts and Agreements
Bibliomation, a nonprofit group serving more than 40 Connecticut libraries, replacing CARL. St. Mark’s School of Texas, in Dallas; replaces Dynix.
Arkansas State University at Newport, replacing Dynix. Southwestern University in Georgetown, Texas, replacing Dynix. University of Saint Mary of the Lake, Mundelein, Illinois, replacing OCLC. Point Park University in Pittsburgh, replacing Millennium. University of Abertay, in Dundee, Scotland, replacing Dynix. Reinhardt College in Waleska, Georgia, replacing EOS.
National and University Library in Pristina, Kosovo. Lawrence University in Appleton, Wisconsin; includes MetaLib and SFX; replacing Sirsi DRA Classic.
Appalachian College Association libraries, 10 colleges in the Appalachian Mountains, replacing nine different systems.
Matanuska-Susitna (Alaska) Library Network.
Dorsey and Whitney, a law firm headquartered in Minneapolis.
The University of Delaware will be the first customer to install VITAL, the digital object repository based on the open source Fedora project. |
Right Sidebar
|
|
