<?xml version="1.0" standalone="yes"?>
<Paper uid="X96-1026">
  <Title>CHINESE INFORMATION EXTRACTION AND RETRIEVAL</Title>
  <Section position="8" start_page="117" end_page="118" type="concl">
    <SectionTitle>
3.5 Schedule
</SectionTitle>
    <Paragraph position="0"> The developer and the Government must devise a schedule that indicates approximate system delivery dates. These delivery dates should be adhered to, to the extent possible, and any slippages should be documented and the cause for the slippage understood.</Paragraph>
    <Paragraph position="1"> The developer should identify any dependencies in the schedule for system deployment. Developers need to manage &amp;quot;requirements crop&amp;quot; and identify potential negative impacts on scheduling initiatives. For technologies being ported to a new language, with a heavy dependence on creation of new resources, developers should track incremental progress to more readily identify problem areas and potential schedule slippage.</Paragraph>
    <Section position="1" start_page="117" end_page="117" type="sub_section">
      <SectionTitle>
3.6 Support
</SectionTitle>
      <Paragraph position="0"> Life cycle support of advanced natural language technology is still beyond the ability of most software centers, which creates an unrealistic support requirement on contractors that focus primarily on research and technology development. The support structure for such a system must be developed and refined as the technology matures in order to be able to handle any future problems.</Paragraph>
      <Paragraph position="1"> While the goal of sharable systems across multiple Government agencies is admirable, all participating agencies must commit to providing support and infrastructure resources to maintain the resulting system within each office. Common system development will provide benefits to the Government in the long term, however it requires substantial initial investment and customer buy in.</Paragraph>
    </Section>
    <Section position="2" start_page="117" end_page="118" type="sub_section">
      <SectionTitle>
3.7 Lessons Learned
</SectionTitle>
      <Paragraph position="0"> The design and development of the demonstration system was a valuable learning experience, which will positively impact the success of future technology efforts. Among the most relevant lessons: * Keep the scope of technology al development efforts small, until an advanced technology is proven to work for a given language.</Paragraph>
      <Paragraph position="1"> * Do not rely on baseline evaluation results to predict the success of a technology effort in a new language. Anticipate that unforeseen challenges of a new language will probably drive system performance down to some degree.</Paragraph>
      <Paragraph position="2"> * Determine reasonable, cost effective means for evaluating new capabilities in existing technologies.</Paragraph>
      <Paragraph position="3">  * Do not rely on evolving standards as a core component of a system.</Paragraph>
      <Paragraph position="4"> Avoid depending on component software still under development without including support for coordination between main system and component developers.</Paragraph>
      <Paragraph position="5"> Include support for any component software as a separate task for project scheduling and budgeting purposes.</Paragraph>
      <Paragraph position="6"> Begin technology development only after the support infrastructure is identified. Develop effective management mechanisms for multiple site coordination with the Government.</Paragraph>
      <Paragraph position="7"> Track incremental progress during the course of system development. This will allow the system integrators and customers to more easily identify potential problem areas.</Paragraph>
      <Paragraph position="8"> Expect that systems shareable across Government agencies require interagency investment far beyond the initial definition of an architecture. The TIPSTER demonstration system allowed us to test implementation of TIPSTER technology in a new language, and gave us a more complete understanding of risks involved in undertaking such an effort. Hopefully this increased understanding will benefit us in future TIPSTER advanced technology transfer efforts.</Paragraph>
    </Section>
  </Section>
class="xml-element"></Paper>