Database Reference
In-Depth Information
NOTES
1. http://www.opencyc.org/ .
2. http://www.ontologyportal.org/ .
3. http://www.loa.istc.cnr.it/DOLCE.html
4. Note even though this is an RDFS property, it is still available for use by OWL ontologies.
5. Obviously, in the case where the two terms have been stated as classes in different
ontologies, then we would need to make them equivalent classes, but this is something
that can be avoided within an ontology.
6. This is the short form of: http://www.opengis.net/rdf# . The OGC geometry ontology
also contains vocabulary for RCC8 topology as well, but in this example we are show-
ing these described in a different imaginary ontology to emphasize how different aspects
can be obtained from different micro-ontologies.
7. This emphasizes the difference between mereology and topology; while it is not unrea-
sonable to exclude a steelworks from being a part of a farm—a mereological exclu-
sion—it is possible that a steelworks could be contained within the extent of a farm,
which is a topological possibility.
8. Note that in Manchester syntax “and” and “that” are used interchangeably. It is the
brackets in the sentence that make it differ from the previous example.
9. Assume here that Enclosed Land is a kind of Landform.
10. It is also worth remembering at this point that whereas subproperties inherit range and
domain restrictions, they do not inherit any OWL restrictions. Although this is some-
what odd, it can be turned to advantage, as we shall see.
11. It should also be remembered that the inverse property will also not be inherited, so this
will have to be explicitly stated for “has direct part.”
12. For now, let us forget about the modification we made to Place by making it a subclass
of Site or Building and just concentrate on the key problem we identified before making
the modification.
13. This definition might look a bit recursive, but in fact it is not. What the statement says
is that if the pattern hasPart o isIntendedFor occurs, this will cause the reasoner to infer
the isIntendedFor property exists.
14. Specifying that a Footprint can be represented by a single point rather than by an area
may seem odd or more bluntly wrong. However, the key here is the word r epresentative :
The hospital obviously has an area associated with its Footprint, and the point is just
used to represent this area in the absence of more precise knowledge about it.
15. In fact, another more scientific definition exists based on gradient and deposition rates
and dependent on the sediment material (Schumm and Kahn, 1972). This definition is
even more convoluted and only presents the same problems. Let's not go there!
16. This is one of the top-level classes of Feature.
17. In turn, this is a kind of Landform.
18. Although habitat and land cover classifications are different, they are nonetheless related
as land cover is a major component of many habitats.
19. Note that in Rabbit “only” means that something must have that property value and can
only have that class or value, whereas in Manchester Syntax to achieve this you have to
use both “some” and “only.”
20. Alternatively, Merean Nature could decide simply to replace its woodland class with the
Merean Maps' class.
21. This is where mm: expands to https://ontology.mereamap.gov.me/ .
22. This is only shown using Manchester syntax as anonymously classes cannot be expressed
in Rabbit.
Search WWH ::




Custom Search