Re: [topicmaps-comment] OpenCyc and Topic Maps: Progress Report

[Martin Bryan <mtbryan@sgml.u-net.com>]

Tom

> This is interesting, thanks.  I just downloaded the hyperlinked version
> (over 13MB), and immediately found what I think is an anomaly.
> "Abbreviation" is said to be
>
>         <rdfs:subClassOf rdf:resource="#ProperNameString"/>
>
> but the comments describing ProperNameString do not agree with those for
> Abbreviation.  An Abbreviation is said to be
>
> "A shortened version of a written word or phrase"
>
> wheras a ProperNameString is said to be
>
> "The collection of all proper names, considered as haracter strings.
>
> I don't see how, with these definitions, Abbreviation can be a subClassOf
> ProperNameString.
>
> You have to figure on there being some errors in a work this massive, but
I
> never expected to hit one so soon.  Or can anyone give a rationale as to
why
> these definitions are compatible?

The whole thing is riddled with errors. Look at the relationships between
the terms dogs and domestic pets. Both are subclasses of Domesticated
Animal, but no relationship is defined between the two terms. A domestic pet
is a "Portable Object" but a dog is not! (If you've ever tried carrying a St
Bernard you might know why this is so, but...!) Domesitc pets are a subclass
of HumanOccupationConstructResident, but dogs are not... Then try looking
for a relationship between domesticated animal food (which for some reason
"includes feed manufactured for horses, cattle, chickens, and other farm
animals, as well as pet food") and domestic pets, etc. Sorry I'm not
impressed on the strength of the DAML definitions (though this may be caused
by limitations in DAML)

Martin Bryan