<?xml version="1.0" standalone="yes"?>
<Paper uid="W97-0909">
  <Title>NLP and Industry: Transfer and Reuse of Technologies*</Title>
  <Section position="7" start_page="75" end_page="75" type="ackno">
    <SectionTitle>
4 Conclusion
</SectionTitle>
    <Paragraph position="0"> The NLP tools described in this paper have been used a number of times to effect the mass-change of on-line textual databases. The cost savings over other methods has been significant (we estimate, for example, that in four years, 20,000 man-hours have been saved over manual methods). By representing core semantic components in Prolog, we expect to minimize the work needed to accommodate radical domain changes in the future, though application-specific manual work must s011 be performed to update the lexicon, modify the grammar, and elaborate new referential semantics.</Paragraph>
    <Paragraph position="1"> These same NLP tools, modified to accommodate primarily lexical difl'erences, a more complicated semantic domain model, and deeper interpretation, have been employed in building NLIs to legacy databases and applications, in a resource-conserving manner.</Paragraph>
  </Section>
class="xml-element"></Paper>