<?xml version="1.0" standalone="yes"?>
<Paper uid="W02-1704">
  <Title>Web References</Title>
  <Section position="3" start_page="0" end_page="0" type="intro">
    <SectionTitle>
2 Dictionaries and XML: Related
</SectionTitle>
    <Paragraph position="0"> work Recent research in lexicology has been focused on two di erent goals: the mark-up process for existing print dictionaries, and the successful construction of machine-readable dictionaries from scratch.</Paragraph>
    <Paragraph position="1"> The rst approach has received more attention in the past. This is partly due to the fact that the transformation of existing print dictionaries into modules for NLP applications promises to be less time-consuming than the construction of a new machine-readable database. Lexicologists agree on the fact that a dictionary entry is inherently hierarchical, i.e., it consists out of atomic elements grouped together within non-atomic elements in a tree-like hierarchy. Many approaches place orthographical and phonological information together in one group, while grammatical information is put in a di erent group. This hierarchical approach also allows to denote scope by inserting information at di erent levels of the hierarchy. Again, information about orthography and phonology generally applies to every facet of the headword and are thus placed high in the hierarchy, while other information might only apply to single de nitions and thus ranks lower hierarchically (Amsler/Tompa, 1988; Ide, V eronis, 1995; Ide et al., 2000).</Paragraph>
    <Paragraph position="2"> A common problem of lexicologists working with print dictionaries is the fact that there is a certain variation between entries in any two given dictionaries or even within the same dictionary. This results in a neccessary trade-o between the descriptive power and the generality of an approach, i.e. to design a SGML application that is both descriptive enough to be of practical value and general enough to accomodate the variation.</Paragraph>
    <Paragraph position="3"> There has been, on the other hand, only little research on machine-readable dictionaries that are not based on print dictionaries. To our knowledge, only (Ide et al., 1993) deals with this issue by reviewing several approaches towards encoding machine-readable dictionaries.</Paragraph>
    <Paragraph position="4"> One of these is the use of text models that apply a rather at hierarchy to mark up dictionary entries. These text models might chie y use typographical or grammatical information.</Paragraph>
    <Paragraph position="5"> Another approach is using relational databases, in which the information contained in a dictionary entry is distributed over several databases. A third approach is based on feature structures that impose a rich hierarchical structure on the data. The authors nally describe an example application that uses feature structures encoded in SGML to set up a machine-readable dictionary. null The papers mentioned above agree on using SGML for the mark-up. We found that their SGML code is, however, in general XMLcompliant. null</Paragraph>
  </Section>
class="xml-element"></Paper>