<?xml version="1.0" standalone="yes"?>
<Paper uid="P04-3031">
  <Title>NLTK: The Natural Language Toolkit</Title>
  <Section position="4" start_page="0" end_page="0" type="metho">
    <SectionTitle>
3 Installing NLTK
</SectionTitle>
    <Paragraph position="0"> NLTK is available from nltk.sf.net, and is packaged for easy installation under Unix, Mac OS X and Windows. The full distribution consists of four packages: the Python source code (nltk); the corpora (nltk-data); the documentation (nltk-docs); and third-party contributions (nltk-contrib). Before installing NLTK, it is necessary to install Python version 2.3 or later, available from www.python.org. Full installation instructions and a quick start guide are available from the NLTK homepage.</Paragraph>
    <Paragraph position="1"> As soon as NLTK is installed, users can run the demonstrations. On Windows, the demonstrations can be run by double-clicking on their Python source files. Alternatively, from the Python interpreter, this can be done as follows:</Paragraph>
    <Paragraph position="3"> 4 Using and contributing to NLTK NLTK has been used at the University of Pennsylvania since 2001, and has subsequently been adopted by several NLP courses at other universities, including those listed in Table 2.</Paragraph>
    <Paragraph position="4"> Third party contributions to NLTK include: Brill tagger (Chris Maloof), hidden Markov model tagger (Trevor Cohn, Phil Blunsom), GPSG-style feature-based grammar and parser (Rob Speer, Bob Berwick), finite-state morphological analyzer (Carl de Marcken, Beracah Yankama, Bob Berwick), decision list and decision tree classifiers (Trevor Cohn), and Discourse Representation Theory implementation (Edward Ivanovic).</Paragraph>
    <Paragraph position="5"> NLTK is an open source project, and we welcome any contributions. There are several ways to contribute: users can report bugs, suggest features, or contribute patches on Sourceforge; users can participate in discussions on the NLTK-Devel mailing list2 or in the NLTK public forums; and users can submit their own NLTK-based projects for inclusion in the nltk contrib directory. New code modules that are relevant, substantial, original and well-documented will be considered for inclusion in NLTK proper. All source code is distributed under the GNU General Public License, and all documentation is distributed under a Creative Commons non-commercial license. Thus, potential contributors can be confident that their work will remain freely available to all. Further information about contributing to NLTK is available at http://nltk.sf.net/contrib.html.</Paragraph>
  </Section>
class="xml-element"></Paper>