Verity Profiler Toolkit Brings Order to Document Glut

Companies drowning in documents can now classify and route them using the new Profiler Kit from Verity Inc. (Sunnyvale, Calif.,

"Eighty percent of business information is stored outside a traditional relational database -- including memos, e-mail, reports, PowerPoint presentations and documentation," says Ron Weissman, vice president of corporate strategy for Verity.

With Verity Profiler, a company can build a set of profiles that consist of business rules against which those documents can be compared. The Profiler can compare up to 100,000 profiles against a stream of documents at a rate of about five documents per second.

The Profiler requires a programmer to build it into a custom application, Weissman explains. As a result, customers can use Verity Profiler to build applications that disseminate documents to the appropriate users, including business intelligence analysis, smart push and smart filing systems. Documents can be generated internally, can be received from external sources or can be culled from legacy document repositories.

"I would expect the kinds of users would be large corporate intranets doing business intelligence -- to let me know every time a competitor issues a press release, for example -- or large e-commerce and online publishing sites that would use it to catalog information as it was created," Weissman says.

Congressional Quarterly, a Washington, D.C.-based service that tracks legislation and other documents concerning the U.S. Congress, is using Verity to help migrate its subscribers from its 11-year-old OpenVMS-based service, called Washington Alert, to a Web-based service, called CQ.COM on Congress.

"We push the limits as far as the searching goes. This is not simple keyword searching," says Larry Tunks, director of IT for Congressional Quarterly, who explains that subscribers perform searches that include combinations of keywords and up to 20 attributes, such as the name of a particular committee or member of Congress.

On the older service, subscribers could receive Night Alerts, a set of documents based on a search performed at a particular time. Now, subscribers can set up a profile to monitor all incoming documents and notify them when a document matching their profile arrives. That notification can now occur either through e-mail or by having a link to the document added to the subscriber’s personal Web page on the Congressional Quarterly site. Previously, only e-mail notification was available, Tunks says.

Because the Profiler monitors documents as they arrive, instead of storing data about each document in a repository and then searching the repository, Congressional Quarterly can now provide near real-time service without overstressing the back-end system, Tunks says. "It [provides] the ability not only to perform the searches, but also to deliver based on the profiles people set up," he explains.