search

amrisi / amr-guidelines

244 stars 87 forks source link

Checker likely error (ordinal-entity) #87

Closed mgeorgescu closed 10 years ago

mgeorgescu commented 10 years ago

Similar to annotating approximately with quantities, where we lifted "approximately" as parent concept, we annotated as follows: However, this is picked up by the checker as ":ord sub-concept should be ordinal-entity".

PROXY_AFP_ENG_20071030_0313.10   Gomshadzehi 's execution is approximately the 240 th hanging in Iran in 2007 according to an AFP count compiled from press reports .

(s / say-01
     :ARG0 (c / count-01
          :ARG0 (p3 / publication
               :name (n / name :op1 "AFP"))
          :ARG1-of (c2 / compile-01
               :source (r / report-01
                    :ARG0 (p / press))))
     :ARG1 (h / hang-01
          :location (c3 / country
               :name (n3 / name :op1 "Iran"))
          :time (d / date-entity
               :year 2007)
          :domain (e / execute-01
               :ARG1 (p2 / person
                    :name (n2 / name :op1 "Gomshadzehi")))
          :ord (a2 / approximately
               :op1 (o / ordinal-entity
                    :value 240))))

Please let us know if this looks ok. Thank you.

uhermjakob commented 10 years ago

Madalina, The way you handle "approximately the 240th" in your AMR looks good to me, even if the checker (still) complains.

This case is probably not super common, but it makes me wonder whether we should revisit Intervals of Ordinals as well to better conform with our "the fuzzy goes on top" principle.

mgeorgescu commented 10 years ago

Sure, that sounds good. Thanks.