Use case: Sub-properties for locn:geometry

Background

• Discussion in the LOCADD CG thread Sub-properties for locn:geometry?, started on 10 Jan 2014, by John Goodwin.

Summary of discussion

The discussion covered the topic of how to model notions as centroid, representative point, bounding box, etc.

Proposed solutions:

1. Define them as sub-classes of class locn:Geometry
2. Define them as sub-properties of property locn:geometry

It was noted (see Sven Schade's mail) that option (1) would increase complexity and limit re-use of the LOCN vocabulary.

In his mail, Sven suggested three criteria to decide about the preferrable option (quoting):

1) What would be a use case for which we would need a feature (or similar) class?

2) Why could this use case not be realized with the more light-weight model that was initially suggested by John?

3) Would such a use case appear outside the ISO TC211 community, which already have a well-established model, the GFM?

The discussion also concerned the lack of the notion of feature in the LOCN vocabulary, and on the possible need to include it, in order to specify its centroid, bounding box, etc. Andrea Perego explained (see his mail) that this notion was intentionally left out - as well as the one of spatial object - (quoting)

[...] due to the intended cross-sector use of the vocabulary. The point made was that such two notions were not supposed to be used outside the geospatial area. So, their possible inclusion was considered more in the scope of a specialised extension to the LOCN voc.

As a consequence, in the LOCN vocabulary, class dcterms:Location covers both the notions of feature and spatial object ().

The group moved towards an agreement about the adoption of option (2) - i.e., modelling centroids, etc. as sub-properties of locn:geometry.

John Goodwin (see his mail) provided some examples of their possible uses (quoting):

os:7000000000037256
  a locn:Location;
  locn:representativePoint os:1234;
  locn:extent os:2458;
  os:1234 a sf:Point;
  os:2468 a sf:Polygon;
  ...  

where sf:* could also be replaced by other vocabularies (e.g., neogeo:Point, neogeo:Polygon).

However, there were still some open questions needed further investigation. These included also the names and the actual semantics for these properties (e.g., centroid vs representative point, extent vs MBR / bounding box)

It was noted that a LOCN primer would help clarify and test the use of these properties.