Finding A Needle In A Haystack: Online Search Tool Organizes Information Into Electronic Books

The tool, called WebBook, is a Georgia Tech Research Institute (GTRI) initiative to build an online library of modeling and simulation knowledge for the defense industry. More importantly, its authors are testing new ways to electronically organize information to make online research in other subjects far more productive, as well.

Content and functionality are the chief concerns of the WebBook project, said GTRI researcher Margaret Loper, who heads a group that has built a WebBook prototype and is working on a production-quality Web site.

"The goal is to meet a need in the modeling and simulation community for access to comprehensive reference material," Loper explained. "A lot of advanced research information has not been captured in an organized fashion, and even established fundamental concepts are not always easy to find. It's good there's so much information out on the Web, but all that information also makes it difficult to find specific answers."

The GTRI research team elected to build a database-driven Web site to store all the modeling and simulation material they could gather. The site is interactive in that experts can submit information to fill gaps in the overall body of information, and they can add new research reports to the database.

The tricky part is figuring out a way to organize and index the information so it's readily accessible to experts and non-experts alike. To start, WebBook data is divided into text blocks or articles of up to 2,000 words each. Each article is defined by a set of attributes or keywords linked explicitly to other articles with the same keywords. Or articles can be linked implicitly by keywords that are similar or related, but not exactly the same. WebBook users enter the appropriate terms into a search engine that selects matching articles and presents them as an electronic book.

"Articles can be organized in a different way to create these taxonomies or books that can be geared to different types of users," said former GTRI researcher Andrew Old. "So, for example, if you're new to the world of modeling and simulation, there's a book geared toward you with the basic concepts you need to know.

"A lot of times, people don't quite know what to search for specifically," Old added. "If we can present these articles as a book, then they don't necessarily need to know -- they can just browse through the book."

The implicit article links will help users narrow or broaden their searches without starting over. "They give the user the ability to navigate to another article or find out about the existence of that article, even if they're not in the same book," he said.

In addition, WebBook gives users the option of selecting additional articles on a subject in either greater or lesser detail, depending on the specificity of their needs.

Although WebBook is devoted to modeling and simulation, its methodology could be widely applied. "You can plug in any type of content to serve any community you want," Old said. "The basic idea is simple: It's getting the right information to people in the easiest format possible."