BASIS TECHNOLOGY RELEASES ROSETTE LINGUISTICS PLATFORM 4.0

Cambridge, Mass. May 10, 2005 - Basis Technology today announced the availability of release 4.0 of its Rosette® Linguistics Platform. The new release features key enhancements to the company’s core technology — new European language support; performance enhancements for Arabic analysis; new dictionaries for Asian languages; named entity extraction; and expanded coverage for language identification.

Rosette release 4 features a comprehensive “base linguistics” layer which now supports ten languages. Services provided by this layer include tokenization, normalization, stemming, part of speech tagging, noun decompounding, sentence boundary detection, and noun phrase analysis. Rosette’s Arabic base linguistics supersedes and improves upon the performance of Basis Technology’s previous Arabic language analyzer. The accuracy and performance of the platform’s European base linguistics modules have also been enhanced; and named entity extraction is now available for Arabic, Chinese, English, French, German, Italian, Japanese, and Spanish, enabling customers to mine unstructured text for key data such as names, places, and dates.

“This new release of Rosette includes features and performance enhancements requested by our customers,” said Steve Cohen, EVP of Basis Technology. “They need to perform deep, accurate analyses of unstructured text in multiple languages, and this release reaffirms our ongoing commitment to their success.”

Rosette release 4 also includes many enhancements to the core Asian language analysis components, including newly created Chinese, Japanese, and Korean lexicons. Basis Technology assembled lexical resources from four top content providers — Appen Pty Ltd (Sydney); City University of Hong Kong; Hangul Research Center (Seoul); and the University of Hawaii — as well as data harvested from open sources. This comprehensive approach gives the company the flexibility to continually evolve the lexicons to meet changing trends in terminology and marketplace needs.

Carl Hoffman, CEO of Basis Technology, said, “Since 1998, Basis Technology has established itself as the world’s top provider of East Asian language analysis software. Our investments in CJK information retrieval technology are unmatched by any of our competitors, and the results are reflected in our market leadership. Today, our software powers more web searches in Chinese, Japanese, and Korean than that from any other linguistics provider, which is why every major web search engine—including America Online, Ask Jeeves, Google, MSN Search, and Yahoo!—are Basis Technology customers with successful businesses in Asia.”

Rosette’s new architecture also enables organizations to build custom applications that can be integrated into the platform.

“Moving to a single platform architecture enables us to quickly prototype and demonstrate custom applications for our government customers,” said Bill Ray, Vice President of Sales at Basis Technology. “We have shown them how Rosette can drive solutions for global name matching; document triage and exploitation; and geospatial fusion.”

Also featured in the latest Rosette release is additional coverage for the language identification module, which automatically identifies the language and encoding of unknown text. New languages include Bahasa Indonesia, Bahasa Malay, Tagalog, and Vietnamese, bringing the total to 92 language/encoding pairs.

Rosette Linguistics Platform release 4 is available for immediate evaluation. For more information, visit www.basistech.com or call 617-386-2000.

About Basis Technology Basis Technology (www.basistech.com) provides software solutions for multilingual text mining and information retrieval applications. The company’s Rosette® Linguistics Platform is a suite of high-performance, highly reliable, interoperable software components designed for applications that analyze and process all the world’s languages.

Top-tier software vendors, content providers, and multinational enterprises rely on Basis Technology’s solutions for Unicode compliance, language identification, multilingual search, normalization, transliteration, and entity extraction. Clients include industry leaders Cisco, Convera, Endeca, FAST, Google, Hewlett Packard, L.L. Bean, Microsoft, Oracle/PeopleSoft, SAS, Siebel Systems, Symantec, and Verity. Customers in the defense and intelligence industry include BBN, CACI, Lockheed Martin, MITRE, Northrop Grumman, and SAIC, as well as US, UK, and Japanese government agencies.

Company headquarters are located in Cambridge, Massachusetts, with branch offices in San Francisco, California; Herndon, Virginia; and Tokyo, Japan.