It's a small world, after all.
In 2002, the Illinois State Library funded a $20,000 grant to develop a knowledge management project among the North Suburban Library System (NSLS) and two other systems that serve all Illinois libraries. The planned knowledgebase would share common documents and the experience and skills of all system staffs. "XML, the DNA of the Knowledge Management Evolution" (Computers in Libraries, pp. 10-15) by Jan Hayes, NSLS assistant director; Christina Stoll, NSLS KM specialist; and Dawne Tortorella, Bell-Cow, Inc. technology consultant, tells how this was accomplished.
The authors state, "Instinctively, we knew that XML would be the key building block of our knowledge management 'life-form.'" But, what else had to happen? Using IXIASOFT's TEXTML Server, which became the cornerstone for the project, providing a repository for storing XML files in their original format, the consulting team developed a customized application. This application enabled bulk file import, file purging based on date range, and search/display of the XML repository data. By interfacing with the TEXTML Server, Altova's Authentic Browser Plug-In could retrieve XML documents from the repository. This plug-in also allowed for on-the-fly editing and entry of events, a crucial part of the knowledgebase.
An XML schema was developed, and Altova's XMLSpy was used to identify the essential data elements in a continuing education (CE) event. This schema would become the basis for not only generating XML files from existing database content, but also for building the searchable indexes in the repository. To extract the common elements of the different systems' CE events, customized programs were designed. This led to the most difficult challenge of the project: adhering to the proper rules of XML coding during the data-extraction process. The final major element of the knowledgebase is XSLT, eXtensible Style-sheet Language Transformation, which plays an important role in the display of query results. To see the complete XSLT used to transform the XML data into HTML, go to http://www.nsls.info/km/CEEvents.xls.
The authors conclude: "In our project, XML proved itself both extensible and adaptable.... Creating XML data was pretty easy. Creating an application based on XML standards throughout was a challenge we are happy we tackled."
Have you ever done an online search for a specific topic and wondered just how thorough your search was? Then you will want to read John December's article "Toolbars Reveal Web Page Relationships" (ONLINE, pp. 37-40). December notes that it is easy to miss valuable Web sites that don't show up in search engine results or in indexes about a specific topic. By utilizing select browser toolbars, searchers can go on the Web, find related and competitor sites, and gain insight into Web-audience preferences.
What can the A9 toolbar (http://toolbar.a9.com), part of the A9.com search engine and a unit of Amazon.com, do to expand online searching? Beyond the usual toolbar functions, you can glean Web site relationship intelligence through the "People Who Visit" option. Through a pop-up menu, you can go to a Web page that lists sites that share common visitors with the site you are currently visiting. As December points out, "The value of these A9-related sites is that you can quickly identify similar or competitor sites." While the list will no doubt include obvious sites, it will also pinpoint sites with less-obvious connections.
What about identifying related sites through advanced analysis of human language? UCmore (http://www.ucmore.com) is one toolbar that provides language analysis. The concept behind UCmore is to use "structured information" to search a list of keyword matches by categorizing the links, thereby greatly reducing the number of matches. UCmore takes an ambiguous search and guides the user through narrowing down the query to attain a much more manageable number of hits. The Google toolbar's "Similar Pages" function works in quite the same way, showing a set of pages that are related to the page being displayed in the Web browser.
While sampling and methodology issues limit toolbar intelligence and some users cringe at the thought of being monitored by toolbar providers, December says that these may be worthwhile trade-offs for the subtle insights obtained into the relationships among Web sites.
When You Wish Upon IR
How can artificial intelligence (AI) help build better information retrieval (IR) systems to better serve the needs of knowledge workers? This is what Peter Jackson, vice president of research and development at Thomson Legal & Regulatory, asks in "Artificial Intelligence and Information Retrieval" (Searcher, pp. 29-33). To help answer this question, Jackson uses the example of ResultsPlus, a system built by West in 2003 that uses a combination of natural-language processing and text-categorization technologies to provide secondary law materials to primary-law research attorneys.
ResultsPlus improves search results through better indexing. As Jackson writes, "Often in information seeking, users search a selected database, not realizing that relevant documents may exist in other databases." This is where document recommendation deviates from collaborative filtering. Rather than recommending sites simply because others have visited them, ResultsPlus takes into account the user's query and certain features from the context of the search and uses the search results from the initial query to expand the new query. In Lexis-Nexis' Entity Indexing Program for news and business text, names are identified and treated as controlled vocabulary terms, searchable in certain document fields when special query syntax is used. The success of this type of information-extraction technology has led to an increased interest in text mining.
Jackson concludes that the success of ResultsPlus, Entity Indexing, and similar applications stems from the fact that each supplements a tool (search engine) with appliancelike features. This bodes well with 21st-century knowledge workers, who want tools for finding relevant documents without having to conduct a humanlike conversation with a computer or depend upon a machine to tell them what information is significant and what is not.
I Am 43 if You Please
How long ago, and yet not so long ago, does being 4 years old seem. As I write this, I am about to turn 43! But instead of pulling out all my gray hairs (then you could call me Baldy, the eighth Dwarf!) and signing up for Botox injections, I prefer to take an entirely different approach. To quote another famous Disney character (originally created by British playwright J. M. Barrie), the eternal boy Peter Pan, "I'll never grow up, never grow up, never grow up. Not me! Not I. Not me! Not me!"
Lauree Padgett is Information Today, Inc.'s manager of editorial services. Her e-mail address is email@example.com.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||computer technology and Internet make this vast universe a small world|
|Date:||Jan 1, 2005|
|Previous Article:||CAS scientists break records.|
|Next Article:||The Skeptical Business Searcher: the Information Advisor's Guide to Evaluating Web Data, Sites, and Sources.|