State consolidates treaty data

By Patricia DaukantasGCN StaffArms control negotiators can quickly search extensive text and image databases at a secure portal recently opened by a State Department bureau.The Treaty Information Portal consolidates numerous databases into a single, secure online storehouse, said Ned Williams, data repository manager for State's Bureau of Verification and Compliance.Capping efforts that started eight years ago, before State assumed arms control monitoring tasks, the portal represents 'a one-stop shop for treaty data,' Williams said.The bureau began operations only last February. It serves as the U.S. watchdog for international arms control treaties and commitments, overseeing compliance with chemical and biological warfare agreements, as well as those covering nuclear weapons and testing.Many of the bureau's employees had worked for the independent Arms Control and Disarmament Agency, but a 1998 restructuring folded them into the State Department. The old agency was abolished in April of last year.In 1992, a committee within the independent disarmament agency began to develop a complete store of information strictly for officials working on arms control, Williams said.Known as the Arms Control Data Repository, the data store used text-based search engines. Its users had to visit one of several agencies in person or ask a database administrator to seek out information. Nevertheless, the repository project laid the foundation for today's Treaty Information Portal.Williams said he started revamping the portal last year out of concern over the year 2000 problem. His team named it the Treaty Information Portal to emphasize its status as the department's single access point to treaty texts, other documents and satellite images.'We wanted something that would be easily recognizable and easily associated with our mission here,' Williams said.The portal gives access to classified and unclassified databases within several State and Defense department agencies, all tied together through the Secret IP Router Network.'That's where we came up with the concept of calling it a portal, because it does provide access to distributed data,' Williams said.Much of the unclassified data can also be found by searching the Web sites of the State and Defense departments, Williams said. But classified information must remain on the classified network.The SIPRnet 'looks just like the Internet, but it's not connected to the Internet,' Williams said. 'It's based on the same type of technology.'Williams' team styled the portal home page to resemble popular Internet portals such as Yahoo.com. The site also has links to SIPRnet mirror images of relevant agencies' home pages on the Web.The unclassified databases are mirrored to the SIPRnet because the networks have to be kept apart, Williams said.The portal software encompasses an array of data storehouses: Lotus Notes documents; Oracle, Basis and Microsoft Access databases; and other document management systems and image libraries. Basis, from Open Text Corp. of Waterloo, Ontario, is a full-text search and retrieval system that some groups at State have used for a long time.For the portal's core software, Williams selected the RetrievalWare search system from Excalibur Technologies Inc. of Vienna, Va.One of the reasons for choosing RetrievalWare, he said, was its ability to search multiple databases and file formats.The portal lets users search by Boolean, concept-based or pattern-matching queries, Williams said. They can do simple Boolean variations, look for exact matches, associate concepts more or less strongly, and limit the number of databases searched.'There's a whole litany of possibilities to expand or restrict your search criteria,' Williams said.RetrievalWare searches out words related to the originally requested keywords, said John Back, Excalibur Technologies' vice president of North American sales. For example, references to 'missile' could be associated with 'bomb' and 'weapon' but not with 'astronaut,' Back said.Say what?RetrievalWare can carry out concept searches in languages other than English, Back said. The software also finds misspelled words and names with multiple spellings, such as that of Libyan leader Moammar Gadhafi.RetrievalWare returns portal queries rapidly, 'I'd say quicker than Yahoo,' Williams said. The content is indexed to speed searching.By the end of last month, all arms control-related agencies were scheduled to have access to the Treaty Information Portal, Williams said.User feedback so far has been positive, he said, because 'everybody is excited about the possibilities of getting information from one site.'The portal software runs on a four-way Compaq ProLiant 6500 server under Microsoft Windows NT at State's Washington headquarters.Over the next 12 months, Williams and his team will continue fine-tuning the classified portal. The group also might be tapped to build a similar portal on State's unclassified network to support knowledge management initiatives, he said.