American Libraries |
||
Site NavigationLeft Sidebar ItemsOnline Features
Follow American Libraries news stories, videos, and blog posts on Twitter.
|
||
My Kingdom for an OPAC
Head of Systems, North Carolina State University Libraries, Raleigh. Column for February 2005 Besides wishing that we had never come up with the arcane—and now anachronistic—term OPAC for online public access catalog, I wish we had one that searched better. I used to be a web OPAC product manager, and in three successive positions at NCSU Libraries, I have failed to give up on web OPAC development. Suffice it to say, I have a rather intense love-hate relationship with the online catalog. After Google’s fast and furious press releases last year about their plan to make library books searchable (AL, Jan., p. 26–27), there’s been a recent flurry of navel-gazing, finger-pointing, and general moaning and groaning about the state of “searching” in libraries. But in Google’s wake I’d like to make a few points just about OPACs, especially after my statement that the ILS is essentially done (AL, Dec. 2004, p. 64–65). That one generated some feedback from a few OPAC developers and innovators, so it’s time to give the OPAC its fair (but critical) shake. Keywords, schmewordsBefore anyone complains about my age relative to library automation, let it be known that I have indeed used a card catalog . . . extensively. I actually grew up with a microfiche catalog at my public library, so even the worst online catalog was a refreshing upgrade. But if I had to blame the stunted development of online catalogs on one thing, it would be the introduction of keyword searching. Not that there’s anything wrong with keywords—they’re great—but our profession’s utter fascination with the ability to keyword search a catalog has kept both librarians and vendors myopic for nearly 25 years. In 2005, there is little agreement between systems regarding nesting, phrase searching, default Boolean operators, stop words, left-anchoring, truncation, or wild-carding. The closest thing we have is the Z39.50 Bath profile, which is not widely adopted. Only a few library systems have relevance algorithms beyond the number of occurrences in the MARC record. In fact, most systems still default to a last-in, first-out display order. And I have not found a patron who is satisfied with any answer as to why a web search engine can return relevant results from four billion full-text websites faster than an OPAC can return a randomly sorted hitlist from one million surrogate records; nor should any patron be satisfied with even a bona fide answer to that question. Who’s to blame?Besides two decades of tunnel vision with the über-feature of keyword searching, librarians could blame vendors. But, to a great extent, vendors develop what librarians ask them to, so part of this is a matter of asking for the wrong things, or not asking for them at all. For instance, the flip side of our keyword glee is a near-obsessive dedication to precise authority searching in online catalogs. Authorities are important, but I would argue only as system-suggested searches (clicking on an author or subject link in a hitlist or record display) after a patron query. Patrons who can construct their own authoritative subject searches are few and far between. Like one of my more famous colleagues and friends, Roy Tennant, I’m all for placing some blame on the MARC record and AACR2 as well. On the other hand, it’s not MARC’s fault that it predated the online catalog and most of the vendors who would come to contend with this strangely formatted record. (Watch eyes glaze over as you try to explain the logic of MARC to any young programmer who grew up with the internet and earned degrees during the advent of XML, SQL, and object-oriented programming.) MARC is what it is, but how vendors have decided to index fields (or not), store the records, and retrieve holdings records for display has stunted the OPAC for too long. Some hope on the horizonCharles Cutter taught us that the first purpose of the catalog was to find a book about which something is known: author, title, subject. Now the purpose is finding things about which nothing is known except a few keywords; or in some cases where nothing at all is known, and patrons prefer to browse through surrogate records, the way one could, ironically, in a card catalog. Several companies have made serious progress in making lemonade from OPAC lemons. Most are taking different approaches to a host of problems in online search and discovery. RedLightGreen—RedLightGreen from RLG (www.redlightgreen.com) indexes over 120 million records, ranked in order of relevance to the search term and based on how many copies of a title exist in libraries. RedLightGreen not only retrieves relevant results, it does so in a timely manner, with a simple-to-use interface. One of its most impressive features is the way it clusters various editions and manifestations of a work into a single display. For example, there are 59 variations of Death of a Salesman; most library systems would scatter these throughout search results, but RedLightGreen renders them in a single display. The system is also set up to link to numerous local catalogs so that users can use RedLightGreen to discover appropriate titles and then look up holdings in a local catalog. AquaBrowser Library—Founded in Amsterdam by MediaLab Solutions (www.medialab.nl), the AquaBrowser Library presents holdings in a totally new way. Instead of requiring exact search terms, the system finds items using associations, stemming, spelling alternatives, synonyms, and translations. This helps patrons who aren’t sure how best to approach a search. The “refine” tool allows patrons to filter queries by format, subject, or genre. Bibliomondo (now owned by ISACSoft) and TLC both have deals with MediaLab to include AquaBrowser in their online search tools. Endeca—Exhibiting one of the most remarkable refinement tools, Endeca (endeca.com) creates a “metarelational index” where result sets are refined on-the-fly to match patron limits of name, subject, format, or locally defined fields: Picture commercial sites that let you limit results by price, model, or delivery time. As Endeca’s Director of Vertical Markets Matt Eichner puts it, “It’s the same technology making a huge splash in the consumer space.” With measurable results like raising IBM’s online revenue by over 50%, Endeca has entered the library space by partnering with TLC, while other pundits and experts plead with system vendors to take a closer look at the technology. Endeca’s other customers include Barnes and Noble, Wal-Mart, Indigo Books, Abebooks, Harvard Business School, the FBI and CIA, and the Library of Congress. Endeca fuses fast searching with guided navigation, and it capitalizes on the sophistication of metadata. This front-end and back-end addition to bibliographic discovery could revolutionize the OPAC. I wanted to try very hard this month to write about something other than “Google’s big announcement.” I had gotten to the point where one more mention of the ubiquitous masters of search would drive me to give up librarianship altogether so that I could become a shepherd (that is, until my boss reminded me that I tend to get the shakes after two hours without an internet connection). Turns out I did not have to look very far to find companies doing some new and interesting things with the old and boring catalog. Contracts and AgreementsEx Libris sales of Aleph 500: Sichuan University Library (China), including SFX and MetaLib, serving a collection of 4.7 million volumes to over 70,000 students. Geac sales of Vubis Smart: Harnett County (N.C.) Public Library, replacing Geac Advance. GIS sales of Polaris: Fourteen libraries in New York comprising the Mohawk Valley Library System and the 35 libraries of the Southern Adirondack Library System (also in New York), replacing DRA Classic; St. Petersburg (Fla.) Public Library System, replacing DRA Classic. AnnouncementsIn news sure to have metasearch vendors scrambling for legal advice, WebFeat has been awarded a patent for its federated search technology. U.S. patent #6,807,539 reads broadly, but is reported to cover WebFeat’s method and technology for managing authentication and session management across a federated group of electronic resources. Innovative Interfaces will offer Baker and Taylor’s Content Café and its extensive database of enriched bibliographic content. The content will be available to Millennium customers through the online cataloging and acquisitions modules. In an effort to meet the demands of libraries requesting individual title selection over bundled content, Ebrary announced in January that it now offers individual title selection from its database of more than 60,000 full-text books, reports, maps, and other digital content. Ebrary will also offer a perpetual-access model on select titles in June. |
Right Sidebar
|
|
