[developers] [itsdb] parse big corpus with itsdb

Stephan Oepen oe at csli.Stanford.EDU
Fri Nov 10 15:04:11 CET 2006


hi again, zhang yi!

> Another use of such an option I can think of is in the coverage test,
> where only the parsability of the sentence is interested. In such
> cases, the creation of the entire parse forest does not seem
> necessary.

indeed, this is a task i had not taken into consideration.  even though
we cannot guarantee that the first packed tree is globally consistent,
once all constraints are applied in unpacking; for a conservatively set
restrictor, i see how a cheap parseability test could be helpful!  this
much we can accomplish without another option, though: PACKING_NOUNPACK
could signal the parser to stop forest creation as soon as there is one
tree.  should i just add that (as an addition to the two tests i had to
touch yesterday already)?

as for another parameter to limit forest construction, i would be in no
way opposed to extra flexibility.  although i might worry slightly that
we are already offering (too) many distinct ways of running the parser,
where some today make better sense than others :-).

                                                           best  -  oe

+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
+++ Universitetet i Oslo (IFI); Boks 1080 Blindern; 0316 Oslo; (+47) 2284 0125
+++     CSLI Stanford; Ventura Hall; Stanford, CA 94305; (+1 650) 723 0515
+++       --- oe at csli.stanford.edu; oe at ifi.uio.no; stephan at oepen.net ---
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++



More information about the developers mailing list