This article orginally appeared in the May-June 1993 issue of Language Industry Monitor IBM is bringing its substantial linguistic offerings to the OEM market. One problem: Big Blue is about a decade too late. From Bruce Carucci, of IBM’s Cary (North Carolina) site, comes word that IBM plans to enter the linguistic software OEM market. Carucci sent a listing of what IBM calls its Natural Language Processing Service, a collection of linguistic functions, including those for spellchecking, hyphenation, morphology, and synonyms. The functions are accessible through a common API and are written in C. Among the languages IBM supports are Afrikaans, Catalan, Danish, Dutch (both prefer- red and modem spellings), English (Australian, UK, and us), Finnish, French (National and Canadian), German (National and Swiss), Greek, Icelandic, Italian, Japanese, Korean, Norwegian (Bokmal and Nynorsk), Portuguese (National and Brazilian), Spanish, and Swedish. The specific support varies for each language; specific information can be obtained from Carucci via email. IBM researchers at various IBM research labs around the world have had the luxury and means to study the (computational) linguistic problems of their respective languages for years. The Cary group has been responsible for consolidating IBM linguistic technology and “packaging” it. Two IBM programs which have made extensive use of the NLP Services are TM/2 and SearchMaster, for example. Whereas previously, IBM would never have licenced such technology to third parties, preferring to keep it in- house, times have indeed changed. Carucci says IBM intends to licence and support NLP Services “very competitively.” IBM, PO Box 6000, MaiLstop TH8/5W/661, Cary, NC, 27511, USA; Tel: +1 9194697280, Email: bcarucci@vnet.ibm.com |