dcsimg
RealTime IT News

Groxis Opens Deep Web to Business, Schools

Groxis, a provider of search clustering and visualization technology, launched the Grokker Research Pilot designed to act as the interface between "deep Web" content, the kinds of information not indexed by regular Web crawlers.

Thanks to business relationships with providers of proprietary content, which often requires a paid subscription, the Grokker Research platform will offer them a permission- or royalty-based model for distributing their content on the Web.

Grokker Research is a template-version of the Grokker search application pre-loaded with content sources for business and education. Grokker delivers search results in the form of "maps," graphical representations of the different topics associated with the query and of the relationships between those topics. The toolset includes built-in email, so that users can quickly share results, and the ability to save searches.

The offering builds on Grokker E.D.U., a search tool announced in December 2004 that was customized for research, rather than Web search. Educational institutions bought subscriptions to information sources but found them underused by students. Groxis worked with them to create a single graphical interface to search across all available content sources.

Adreanne Radonich, director of marketing of operations for Groxis, said working with institutions like Stanford University built relationships with many providers of private content. "Through Stanford and our other customers, we're building relationships with content providers who see the tool as way to organize and display their data in new and meaningful ways."

Businesses and institutions pay for an enterprise-wide annual subscription license. The license includes the basic Grokker search features and connectors to sources on the Web such as Amazon.com and eBay , plus a handful of premium content providers. The software will limit access to premium content to users who have subscription rights.

Grokker's retrieval engines use XML technology to extract data from digital content sources. Radonich said data providers were very willing to put their information into XML format so that the search tool can "grok" it.

Groxis will write data connectors to tap into more kinds of subscription content that customers want to display in the Grokker interface. "Once we have the connectors written, it will be easy to change the parameters based on different licenses that end customers may have," Radonich said.

For example, the company worked with Sun Microsystems to create a unified, graphical interface for premium content in the companys corporate research library.

The library offered users EBSCO information services (a catalog of electronic titles), the IEEE Xplore electronic library and a repository of research-related e-mail. While each was accessible to corporate users through the intranet or the Web, different interfaces and protocols were a barrier for users.

The SunLibrary Grokker, which went live in December 2004, allows users to search multiple data sources at once, as well as other search engines' indexes. Search results are grouped by topics rather than by source.

The product announcement marks a strategic move by the company away from selling a desktop application. The move began in May, with the release of a Web search service at www.grokker.com powered by Yahoo's search technology. The site hooks into Yahoo's search API and applies Groxis' clustering and visualization technology to the search results.

People were interested in the Grokker tool, but not necessarily in maintaining a piece of desktop software," Radonich said.

Eventually, she said, Groxis plans to offer public access to deep Web content, both free information that hasn't been accessible to Web search and paid content. The company wants to enable pay-as-you go access to subscription content.

"If we could allow it to happen on the site for a minimal fee or for free, that puts us in a really interesting position," Radonich said. "There is a way to work with directly with publishers to get their content out to the everyday researcher, rather than just to people who have enterprise licenses to the content."