<?xml version="1.0" standalone="yes"?>
<Paper uid="W04-0601">
  <Title>Techniques for Text Planning with XSLT</Title>
  <Section position="3" start_page="0" end_page="0" type="intro">
    <SectionTitle>
2 Systems
2.1 COMIC
</SectionTitle>
    <Paragraph position="0"> COMIC1 (den Os and Boves, 2003) is an ongoing project investigating multimodal dialogue systems. The demonstrator adds a dialogue interface to a CAD-like application used in bathroom sales situations to help clients redesign their rooms. The input to the system includes speech, handwriting, and pen gestures; the output combines synthesized speech, a &amp;quot;talking head&amp;quot; avatar, and control of the underlying application. Figure 1 shows screen shots of the avatar and the bathroom-design application.</Paragraph>
    <Paragraph position="1"> COMIC produces a variety of output, using its full range of modalities. In this paper, we will concentrate on the textual content of those turns in which the system describes one or more options for  decorating the user's bathroom, as in the following description of a set of tiles: (1) Here is a country design. It uses tiles from Coem's Armonie series. The tiles are terracotta and beige, giving the room the feeling of a Tuscan country home. There are floral motifs on the decorative tiles.</Paragraph>
    <Section position="1" start_page="0" end_page="0" type="sub_section">
      <SectionTitle>
2.2 OpenCCG
</SectionTitle>
      <Paragraph position="0"> The OpenCCG realizer (White and Baldridge, 2003) is a practical, open-source realizer based on Combinatory Categorial Grammar (CCG; Steedman, 2000). It employs a novel ensemble of methods for improving the efficiency of CCG realization, and in particular, makes integrated use of n-gram scoring of possible realizations in its chart realization algorithm (White, 2004a; White, 2004b). The n-gram scoring allows the realizer to work in &amp;quot;anytime&amp;quot; mode--able at any time to return the highest-scoring complete realization--and ensures that a good realization can be found reasonably quickly even when the number of possibilities is exponential. null Like other realizers, the OpenCCG realizer is partially responsible for determining word order and inflection. For example, the realizer determines that also should preferably follow the verb in There are also floral motifs on the decorative tiles, whereas in other cases it typically precedes the verb, as in It also has abstract shapes. It also enforces subject-verb agreement, e.g., between are and motifs, and it and has, respectively. Less typically, in COMIC and FLIGHTS, the OpenCCG realizer additionally determines the type of pitch accents, and the type and placement of boundary tones, based on the information structure of its input logical forms.</Paragraph>
    </Section>
  </Section>
class="xml-element"></Paper>