<?xml version="1.0" standalone="yes"?>
<Paper uid="W04-1710">
  <Title>The Syntax Student's Companion: an eLearning Tool designed for (Computational) Linguistics Students</Title>
  <Section position="8" start_page="0" end_page="0" type="concl">
    <SectionTitle>
7 Perspectives and conclusions
</SectionTitle>
    <Paragraph position="0"> One could think of many other features that would probably make the program even more useful for learning. We only mention a few and we hope that OpenSource contributions will extend the list.</Paragraph>
    <Paragraph position="1"> A key aspect of this kind of support tool certainly lies in the nature of the feedback that is provided to students. We have already said that the mode for deflning exercises will allow the teacher to specify possible wrong solutions and to associate them with an appropriate correction. An interesting extension would be a mode where students could send the results of their exercise session (possibly containing a series of coherent exercises) to a supervisor by 14In the latter case, it will be possible to specify that the trees be developed with an upright lefthand branch and sloping righthand ones, as this layout is used in some textbook and is therefore more familiar to students using them.</Paragraph>
    <Paragraph position="2"> email from the program. Then, the annotated corrections of exercises could feed a database and be reused in subsequent unsupervised exercises. We think that there is indeed much to be gained from past corrections, as shown in the research on vicarious learning using past dialogues between learners and their teachers (Cox et al., 1999), which, incidentally, was also based on the teaching of syntax.</Paragraph>
    <Paragraph position="3"> The range of topics covered by the program could be extended. The learning of syntax could probably be supported by the integration of parsers, which could be of particular interest to computational linguistics students (see e.g.</Paragraph>
    <Paragraph position="4"> (Meurers et al., 2002; van Halteren, 2002)).</Paragraph>
    <Paragraph position="5"> The integration of generators would also allow students to inspect the productions of their grammars to attempt to identify why they could overgenerate. Furthermore, we would like to reuse what already exists for the morphological analysis of words in terms of in ections and derivations, as well as for compositional semantic analysis.</Paragraph>
    <Paragraph position="6"> The program we have presented puts a particular emphasis on its central users, who are students in (computational) linguistics. Initial evaluation has shown that this kind of support was very welcome by the learners' community, and we hope that it will be more widely adopted by the teachers' community in its new version that attempts to reduce known limitations. We look forward to new developments in the fleld of research in computer-assisted learning, and in particular on methodologies for the evaluation of systems.</Paragraph>
  </Section>
class="xml-element"></Paper>