[developers] characterization
Ben Waldron
benjamin.waldron at cl.cam.ac.uk
Thu Nov 17 11:38:48 CET 2005
Sorry, I'm off work ill at the moment. I'll get back to you fully when
I'm better, but very briefly: things are not yet set up such that
(character) positions are passed from ChaSen into the LKB when using the
JACY grammar, and I'll have to take a look to see what might be causing
your stack overflow.
(Aside: I failed to mention in my previous email that users of the
internal LKB preprocessor need to call the
x-read-preprocessor/x-preprocess functions in place of
read-preprocesor/preprocess in order to pass character positions to the
LKB parser.)
- Ben
Francis Bond wrote:
>G'day,
>
>Thanks for working on the characterization. I had a little play with
>it. It (lkb::*cvs-version* "$Date: 2005/10/25 16:56:39 $") works
>when exporting from [incr tsdb()], but when I try and do it
>interactively it almost always fails with the following warning:
>Error: Stack overflow (signal 1000)
> [condition type: SYNCHRONOUS-OPERATING-SYSTEM-SIGNAL]
>
>This makes testing tedious, as I have to restart lisp each time.
>
>Also, the cfrom/cto values seem to be token positions. I thought the
>consensus was that they should be characters. Or if you want
>character positions do you have to give them explicitly in the input
>in some way?
>
>
>--
>Francis Bond <www.kecl.ntt.co.jp/icl/mtg/members/bond/>
>NTT Communication Science Laboratories | Machine Translation Research Group
>
More information about the developers
mailing list