<?xml version="1.0" standalone="yes"?>
<Paper uid="P91-1013">
  <Title>LR RECURSIVE TRANSITION NETWORKS FOR EARLEY AND TOMITA PARSING</Title>
  <Section position="12" start_page="103" end_page="104" type="concl">
    <SectionTitle>
8. CONCLUSION
</SectionTitle>
    <Paragraph position="0"> We have introduced BLR(0), a simple bottom-up LR RTN-based CF parsing algorithm. We explicitly expand grammars to RTNs, and only then construct our parsing algorithm. This intermediate step eliminates the complex algebra usually associated with parsing, and renders more transparent the close relations between different parsers.</Paragraph>
    <Paragraph position="1"> Earley's algorithm is seen to be fundamentally an LR parser. Earley's propose expansion step is a recursion analogous to our Follow-Set traversal of the RTN. By explicating the LR-RTN graph in the computation, no other complex data structures are required. The efficient merging is accomplished by using an option available to BLR(0): merging parse nodes into equivalence classes.</Paragraph>
    <Paragraph position="2"> Tomita's algorithm uses the cached LR Follow-Set option, in addition to merging. Again, by using the RTN as a concrete data structure, the technical feats associated with Tomita's parser disappear. His shared packed forest follows immediately from our merge option. His graph stack and his parse forest are, for us, the same entity: the shared parse tree. Even the LR table is seen to derive from this parsing activity, particularly with incremental construction from the RTN.</Paragraph>
    <Paragraph position="3">  Bringing the RTN into parsing as an explicit realization of the original grammar appears to be a conceptual and implementational improvement over less uniform treatments.</Paragraph>
  </Section>
class="xml-element"></Paper>