<?xml version="1.0" standalone="yes"?>
<Paper uid="H93-1075">
  <Title>A SIMULATION-BASED RESEARCH STRATEGY FOR DESIGNING COMPLEX NL SYSTEMS*</Title>
  <Section position="7" start_page="374" end_page="374" type="concl">
    <SectionTitle>
5. FUTURE DIRECTIONS
</SectionTitle>
    <Paragraph position="0"> The long-term goal of the present research method is to support a wide spectrum of advance empirical studies on interactive speech, pen, and pen/voice systems under different circumstances of theoretical and commercial interest. Future extensions of the present simulation research are under way to examine issues relevant to multilingual and other multiparty applications \[13\]. In addition, a taxonomy of tasks is being developed in order to establish a more analytical basis for distinguishing when findings do or do not generalize to qualitatively different domains, such that future work need not approach each new application as an unknown entity. Efforts also are under way to define the important dimensions of system interactivity, such as feedback characteristics and error resolution strategies, as well as their impact on human-computer interaction. Finally, in addition to providing proactive guidance for system design, a further aim of this simulation research is to yield better information about the range of preferred metrics for conducting performance assessments of future NL systems, including their accuracy, efficiency, learnability, flexibility, ease of use, expressive power, and breadth of utility.</Paragraph>
  </Section>
class="xml-element"></Paper>