<?xml version="1.0" standalone="yes"?>
<Paper uid="W06-2806">
  <Title>Interpreting Genre Evolution on the Web: Preliminary Results</Title>
  <Section position="6" start_page="36" end_page="37" type="concl">
    <SectionTitle>
4 Conclusions and Future Work
</SectionTitle>
    <Paragraph position="0"> The study shows a composite picture of the perception of the genre repertoire on the Web.</Paragraph>
    <Paragraph position="1"> This picture focuses on recent genres only, overlooking those more based on paper genres because, in our opinion, this hot area can reveal more about the dynamics behind genre evolution.</Paragraph>
    <Paragraph position="2"> Preliminary findings coming out from this study confirm the initial hypothesis and show that users' perception can be divided into three ranges. These three ranges can be interpreted in terms of genre evolution: high perception for the most stable and acknowledge genres; medium perception for emerging genres, not fully acknowledged by the majority or still unstable, and finally low perception for the highly ambiguous genres (for different reasons). Some of the new web genres can be unambiguously perceived (for example, personal home page, eshop, corporate home page, FAQs and search page).</Paragraph>
    <Paragraph position="3"> Web users can also handle a certain degree of granularity, for example by distinguishing a personal home page from a corporate home page, but the boundary between academic home pages and organizational home pages is still too fuzzy for them.</Paragraph>
    <Paragraph position="4"> The approach to the web as a genre repertoire in evolution and these preliminary findings can turn out to be useful when building web genre palettes or when designing new genre identification experiments.</Paragraph>
    <Paragraph position="5"> Future work includes the computation of agreement coefficients. K statistic is largely used  but still controversial and mostly used for measuring the agreement of two or three raters. Two new interesting measures to assess users' recognition of web page genres were used by Rosso (2005: 109 ff.), but their full interpretation is still under study. The challenging follow up of these preliminary results is to find an objective coefficient of agreement applicable for 135 raters that can choose among 23 categories to classify</Paragraph>
  </Section>
class="xml-element"></Paper>