Print

Print


On 5/13/05, Syd Bauman <[log in to unmask]> wrote:
> > > That guy is <persName key="ART">Arthit</persName>.
> > > His nickname is <name type="person" key="ART">Art</name>.
>
> Lou is correct, the usage is fine, but I think hammering home a point
> he raised earlier is worthwhile. All of the "naming" elements
> (<name>, <persName>, <placeName>, <orgName>, and to some extent even
> <rs>) are used to encode the *name* of a thing, not the thing itself;
> the signifier, not the signed.

Oh, you save me :)
Actually I'm about to ask this question since yesterday,
but has a problem on how I could ask ... I mean, arrange a question.
But you already did it for me, thank you :)


> signified object. key=, however, may be used to specify an identifier
> for the object, thus disambiguating it from other objects with the
> same name. So
>
>   <name type="ship" key="CV00006">USS Enterprise</name>, launched 1938
>   <name type="ship" key="CVN0065">USS Enterprise</name>, launched 1962
>   <name type="ship" key="NXC0001">USS Enterprise</name>, launched 2151
>   <name type="ship" key="NCC1701">USS Enterprise</name>, launched 2245
>
> This identifier may be a key into a database; the corresponding
> record may have all sorts of useful information about the object that
> was named, including (e.g.) a normalized version of its name.

Thank you again.
At this point, it is clear for me that in order to use the "naming" elements
for coreference resolving application, we always need "key" attribute.

cheers,
Art

--
:: Art / Arthit Suriyawongkul
:: http://siit.net/members/art/  |  http://bact.blogspot.com/

--
Thai Wikipedia -- Free Encyclopedia
http://th.wikipedia.org/