Re: why Vehicle subClassOf Product ? (also: Commercial, Economic)

On 27 Mar 2015, at 14:43, Dave Caroline <dave.thearchivist@gmail.com> wrote:

> I think this particular thread exposes my basic worries about mark up
> and the automated reading of it.
> 
We are long beyond the naive idea of consuming the raw data from markup. The question is whether we can create sweet spots the enable complex combinations of deterministic and probabilistic processing steps. This is not Semantic Web 1.0 style ;-)

> Follow some arbitrary rule someone thought up for one or few occasions
> which does not fit all situations or should one use the old method
> where good visible text the user reads explains the item/s better.
> Or extend something that later becomes deprecated or what...

We are in a learning process, but it is widely agreed that trying various conceptualizations is better than waiting for a better conceptualization in the distant future.
The Semantic Web community has brought up wonderful methodologies and tools for building Web ontologies, but never created the latter.

> I saw someone saying once an engine was subclass of vehicle, wrong
> when an engine used on a farm or building.
> e.g. http://www.middleton-leawood.org.uk/leawood/images/pumphouse/index.html
> in this case the "product" is a visitor attraction with opening times
> conating a few things like steam engine and boilers.

Making an engine a subclass of a vehicle is almost always wrong and that is known in the research communities since the 1983 paper by R. Brachman "What IS-A Is and Isn't: An Analysis of Taxonomic Links in Semantic Networks", available from
 
    http://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=1654194

The only case I know where this might be correct modeling is under a "topic semantics", e.g. when you are organizing things that are related to a certain topic (e.g. books about engines might be considered books about vehicles).


Martin

Received on Friday, 27 March 2015 14:12:02 UTC