What do you get when you combine an easy-to-use interface with a database of billions of records and almost guaranteed results? If you guessed Google, you'd be right. But wouldn't it be great if the answer were your library?
What would it take for libraries to provide their patrons with the same kind of success in finding information using our resources as they achieve with Google, and as quickly? The answer is not that complicated: be more like Google, then move beyond Google. And with the technology available right now, it is not an impossible goal.
"People often ask me: 'What's the most important thing I should do if I want to make sure my web site is easy to use?' The answer is simple. It's not 'Nothing important should ever be more than two clicks away,' or 'Speak the user's language,' or even 'Be consistent.' It's 'Don't make me think!'" (Krug 2000, 11)
Substitute "library" for "web site" then add the word "unnecessarily" to the end, and Krug's plea for simplicity and transparency might seem to echo the typical library user's wishes. After all, libraries are complex environments, and librarians have spent decades developing complex structures to describe the information contained within the library's walls, structures that have not adapted all that gracefully to virtual environments. For the library user, the desire to find information and find it fast is paramount, and libraries are often barriers to this goal, unlike the web.
Google, one of the most recognized search engines, does not make users think: its purpose is immediately discernable by the uncluttered prominence of its search box, and even novice searchers are able to obtain useful results. But Google is deceptively simple in its appearance; behind the single search box is a wealth of features that allow for complex searching. In fact, Google offers many of the search features enumerated by Karen Schneider in a recent article about online catalog deficiencies: relevance ranking, word stemming (searching for alternate word endings, e.g., search, searches, searching), field weighting (applying more importance to keywords that appear in the title or first few words of the page), spell-checking, refining original queries (that is, searching within results), support for query operators (for example, + for "required" and - for "not"), Boolean searching (including quoted phrase searches), and in-line query limiters (e.g., site:edu) (2006b).
Also lurking behind the single search box is the Google equivalent of federated, or cross-database, searching. Google offers a variety of search tools for specialized resources such as street maps, images, news headlines, and definitions, each of which can be searched separately. But each of these types of information sources can also be retrieved through Google's main search box; for example, typing a street address into the Google search box will return a link to a street map, along with any related results.
Libraries, in contrast to Google, are not easy to use. A library's web site is the doorway into its collections, both print and electronic, yet most library web sites present users with a number of barriers to finding relevant information, barriers that are both imposed by vendors and created the libraries themselves.
The legacy of paper-based research systems is still evident in today's virtual library environments. Many library web sites still make a distinction between databases for accessing materials owned by the library (e.g., online catalogs, digital collections) and those provided by commercial vendors (e.g., online article indexes, e-journal collections); access to any of these resources can require navigating through two or three layers of a web site. But in the online environment, this distinction between ownership and access is somewhat dubious: it is not unusual for an online catalog to contain records of items, such as web sites, not owned by the library, and commercial databases include records for resources available in the library's print collection. From the user perspective, obtaining relevant information is paramount; whether it comes from the online catalog, the web, or a subscription database is less important.
One technology that allows libraries to emulate the "don't make me think" environment provided by Google is federated searching. With federated searching, users can search articles indexes, the library's online catalog, and even locally developed or web databases all at the same time. Access to the search box can become a feature of a library's home page, eliminating the need to navigate through multiple layers of a web site.
Most federated search software is available from commercial vendors, either as a locally installed option or hosted service; most integrated library system vendors offer a federated search option, as do other vendors such as Ex Libris (MetaLib), webFeat, Inxight, Endeca, and Serials Solutions. To date there appear to be only two open source solutions: Keystone's Digital Library Suite (http://www.indexdata.dk/keystone/) and dbWiz, part of the reSearcher suite (http://researcher.sfu.ca/).
The features offered by federated search tools can vary widely, but can include all of the advanced search features of Google listed above, along with others such as de-duping and sorting of results (including sorting by the database from which the results were retrieved). Like any search tool, federated searching does not eliminate the need for the researcher to select, evaluate, summarize, and integrate sources, but it does remove some decision-making about which source to search. Federated searching represents a big step forward for libraries, but it is not the whole story there is at least one more chapter to explore.
The single search box was one of the most frequently requested features by students participating in a recent study conducted by librarians at Oregon State University, but the students also had two other requests: help them with the steps to narrow down their search and suggest search terms (Mellinger and Nichols 2006). Many library databases, such as online catalogs and article indexes, offer options for narrowing or limiting search results and employ controlled vocabularies that can assist the user in discovering appropriate search terms. But these are not always readily apparent to the library user, who must first know first that they exist, then where to find them in the interface.
But once again, we can look to the web for solutions. Faceted browsing, or search clustering, allows the user to narrow down their search results contextually, revealing additional search terms in the process. In faceted browsing, the clusters are revealed alongside the search results, enabling the researcher to see the context(s) surrounding the search topic and offers "one-click" access to a more focused search result set.
Clusty (http://clusty.com/), a metasearch engine for the web, uses Vivisimo's clustering technology to display results. KartOO (http://www.kartoo.com/), another metasearch engine, combines clustering technology with a visual map of the search results that diagrams the relationships between the clusters. EBSCO recently introduced clustering in its database search results, and now offers a third search interface: the visual search.
There are a number of commercial vendors offering search clustering products, but relatively few libraries have adopted them, as the technology is still quite new. Endeca Systems' product was selected by North Caroline State University. In addition to listing clusters by various facets, including topic, genre, author, and language, users can browse results by Library of Congress class.
AquaBrowser, by Medialab Solutions, was developed specifically for libraries and has the most widespread adoption; most of its customers are public libraries in the United States and The Netherlands. One of the features of AquaBrowser is the visual map of results. Clusters are displayed textually, but a diagram also maps the relationships between terms. When a user clicks on a term in the map, the records listed in the results are refreshed to show the new result, and the diagram is rearranged to show the selected term within its new context.
These two technologies federated searching and clustering when used in concert can vastly improve the research experience for our library users by introducing both simplicity through the single search box and transparency revealing the context within which the searcher's topic is situated. Now if we can just find a product that does it all, and does it well.
Krug, Steve. 2000. Don't Make Me Think: A Common Sense Approach to Web Usability. Indianapolis: New Riders Publishing.
Mellinger, Margaret, and Jane Nichols. 2006. Subject Search Disconnect: Or, How Do Our Users Want to Search for Subject Information? Paper presented at the Oregon Library Association Annual Conference, Salem, OR.
Scheider, Karen. 2006a. How OPACs Suck, Part 1: Relevance Rank (Or the Lack of It). ALA TechSource. http://www.techsource.ala.org/blog/2006/03/how-opacs-suck-part-1-relevance-rank-or-the-lack-of-it.html (accessed April 12, 2006).
Schneider, Karen. 2006b. How OPACs Suck, Part 2: The Checklist of Shame. ALA TechSource. http://www.techsource.ala.org/blog/2006/04/how-opacs-suck-part-2-the-checklist-of-shame.html (accessed April 12, 2006).
© 2006, colleen bell about
transplantedgoose.net / words