There is also a U.S.
government-sponsored program on information extraction
that has focused much attention on the problem of aggregating textual references (not just names, but also referential descriptions and pronouns) into entity-level representations, and normalizing aspects of their meaning (see http://www.nist.gov/speech/tests/ace/). Our work explores the potential usage of gazetteers for NLP purposes in some detail and addresses issues associated with normalizing the meaning of place name references using the gazetteers.