<?xml version="1.0" standalone="yes"?>
<Paper uid="W03-0802">
  <Title>WHAT: An XSLT-based Infrastructure for the Integration of Natural Language Processing Components</Title>
  <Section position="4" start_page="2" end_page="2" type="relat">
    <SectionTitle>
3.5 Related Work
</SectionTitle>
    <Paragraph position="0"> As argued in Thompson and McKelvie (1997), standoff annotation is a viable solution in order to cope with the combination of multiple overlapping hierarchies and the efficiency problem of XML tree modification for large annotations.</Paragraph>
    <Paragraph position="1"> Ide (2000) gives an overview of NLP-related XML core technologies that also strives XSLT.</Paragraph>
    <Paragraph position="2"> We adopt the pragmatic view of Carletta et al. (2002), who see that computational linguistics greatly benefits from general XMLification, namely by getting for free standards and advanced technologies for storing and manipulating XML annotation, mainly through W3C and various open source projects. The trade-off for this benefit is a representation language somewhat limited with respect to linguistic expressivity.</Paragraph>
    <Paragraph position="3"> NiteQL (Evert and Voormann 2002) can be seen as an extension to XPath within XSLT, has a more concise syntax especially for document structure-related expressions and a focus on timeline support with specialized queries (for speech annotation). The query language in general does not add expressive power to XSLT and the implementation currently only supports Java XSLT engines.</Paragraph>
    <Paragraph position="4"> Because of unstable standardization and implementation status, we did not yet make use of XQuery (Boag et al. 2002). XQuery is a powerful, SQL-like query language on XML documents where XPath is a subset rather than asublanguageasinXSLT.</Paragraph>
    <Section position="1" start_page="2" end_page="2" type="sub_section">
      <SectionTitle>
3.6 Advantages of WHAT
</SectionTitle>
      <Paragraph position="0"> wrapper code is relatively small, WHAT can easily be ported to any programming language for whichanXSLTengineexists.Currently,WHAT is implemented in Java (JAXP/Xalan) and C/C++ (Gnome libxml/libxslt). Through libxml/libxslt, it can also easily be ported to Perl, Python, Tcl and other languages * easy to extend to new components/DTDs. This has to be done only once for a component through XSLT query library definitions, and access will be available immediately in all programming languages for which a WHAT implementation exists * powerful (mainly through XPath which is part of XSLT).</Paragraph>
      <Paragraph position="1"> WHAT can be used * to perform computations and complex transformations on XML annotation, * as uniform access to abstract from component-specific namings and DTD structure, and * to exchange results between components (e.g., to give non-XML-aware components access to information encoded XML annotation), * to define application-specific architectures for online and offline processing of NLP XML annotation.</Paragraph>
    </Section>
  </Section>
class="xml-element"></Paper>