<?xml version="1.0" standalone="yes"?>
<Paper uid="W04-0213">
  <Title>The Potsdam Commentary Corpus</Title>
  <Section position="5" start_page="0" end_page="0" type="concl">
    <SectionTitle>
4 Conclusions
</SectionTitle>
    <Paragraph position="0"> The PCC is not the result of a funded project.</Paragraph>
    <Paragraph position="1"> Instead, the designs of the various annotation layers and the actual annotation work are results of a series of diploma theses, of students' work in course projects, and to some extent of paid assistentships. This means that the PCC cannot grow particularly quickly. After the first step towards breadth had been taken with the PoS-tagging, RST annotation, and URML conversion of the entire corpus of 170 texts12, emphasis shifted towards depth. Hence we decided to select ten commentaries to form a 'core corpus', for which the entire range of annotation levels was realized, so that experiments with multi-level querying could commence. Cur- null diploma thesis work of David Reitter (2003), which deserves special mention here.</Paragraph>
    <Paragraph position="2"> rently, some annotations (in particular the connectives and scopes) have already moved beyond the core corpus; the others will grow step by step.</Paragraph>
    <Paragraph position="3"> The kind of annotation work presented here would clearly benefit from the emergence of standard formats and tag sets, which could lead to sharable resources of larger size. Clearly this poses a number of research challenges, though, such as the applicability of tag sets across different languages. Nonetheless, the prospect of a network of annotated discourse resources seems particularly promising if not only a single annotation layer is used but a whole variety of them, so that a systematic search for correlations between them becomes possible, which in turn can lead to more explanatory models of discourse structure.</Paragraph>
  </Section>
class="xml-element"></Paper>