<?xml version="1.0" standalone="yes"?>
<Paper uid="C90-2034">
  <Title>To Parse or Not to Parse: Relation-Driven Text Skimming</Title>
  <Section position="7" start_page="0" end_page="0" type="concl">
    <SectionTitle>
6 Conclusion
</SectionTitle>
    <Paragraph position="0"> Natural language text processing has reached a point where efficiency is a reM issue. While it might be possible to design fundamentally faster text processing algorithms, a more fruitflfl approach in the near term is to try to eliminate much of the wasted processing that is done in parsing text. This does not mean using less syntactic knowledge, but does mean less syntactic processing. Our approach uses the identification of linguistic relations as a driver for producing conceptual information while eliminating some of the detail of parsing. This approach has been successflfl in producing a program that operates ahnost an order of magnitude faster in nm!tiple domains without major effect on the design or accuracy of the system.</Paragraph>
  </Section>
class="xml-element"></Paper>