[developers] variable type mapping as part of VPM
yzhang at coli.uni-sb.de
Fri Aug 8 08:41:55 CEST 2008
thanks for pointing this out. i will work on it next week.
On Thu, Aug 7, 2008 at 10:44 PM, Stephan Oepen <oe at ifi.uio.no> wrote:
> hi zhang yi,
> i am about to check in changes to the (Lisp) VPM code, to finally have
> the mappings from grammar-internal to MRS variables be part of the VPM
> too (such that, among other things, the mapping can be bidirectional).
> adding these to the ERG `semi.vpm' (see attachment), cheap complains:
> Invalid line while reading vpm.
> the intention is that variable type mappings take the same form as the
> variable property mappings, but they occur outside a block of property
> name correspondencnes, i.e. before any X : Y statements.
> conceptually, this extension of the VPM makes the following obsolete, i
> mrs-event-type := event.
> mrs-event-or-index-type := individual.
> mrs-deg-ind-type := deg-ind.
> mrs-non-expl-ind-type := non_expl-ind.
> mrs-handle-type := handle.
> mrs-ref-ind-type := ref-ind.
> mrs-non-event-type := non_event.
> in the LKB, the old way of mapping variable types will remain available
> (for backwards compatibility, at least for some time), but for PET i am
> tempted to suggest we build the new MRS code without such legacy (other
> than the ERG, no grammars support native MRS mode in PET yet, i think)?
> either way, would you have time to extend the PET VPM implementation?
> best wishes - 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 ifi.uio.no; oe at csli.stanford.edu; stephan at oepen.net ---
More information about the developers