Restrictions on how ASD grammars should be written to be used with
ASDParser:
Copyright 1995, 1999, 2008
James A. Mason
The parsing algorithm that is
used by ASDParser imposes a few restrictions
on
how ASD grammars should be
written that are to be used with the parser:
- Dummy nodes in the grammar should never be initial nodes. The parser will
only use initial nodes in the grammar that correspond to actual words
in the given utterance.
- For each initial node in
the grammar, there should be at least one final node reachable from it.
Otherwise, the parses starting at that initial node can't possibly
succeed.
- Since dummy (null) nodes in the grammar don't match any words in
the
utterance being parsed, there must not be cycles of dummy nodes in the
grammar. Otherwise the parser can get into an unending loop in
such a
cycle.
ASD
notes
and documentation
ASD home page
last modified 2010 Aug 21