Hi Gregor,
It is indeed practical to have ids for very specific localities, not just higher geographic features. But I think you misunderstood me. The localityID can apply to any level of granularity, but it should always be the id for the whole of the Location, not some arbitrary part of it, such as the stateProvince in one case and country in another.
In other words, if a country, stateProvince, and county were given with a localityID, that localityID should apply to the combination of the three (in other words, the county, within that stateProvince within that country). If a locality were also given, then the id shoould be for that locality within that county within that stateProvince within that country. The id could even apply to individual georeferences within named places.
John
On Tue, Aug 18, 2009 at 12:08 PM, Gregor Hagedorng.m.hagedorn@gmail.com wrote:
The extremely important thing that may be slipping through the cracks in this discussion is that the locationID should be the identifier for the whole content of the Location part of the record, not just one part of it.
In other words the following combination would be appropriate:
locationID="rs.tdwg.org/ontology/voc/GeographicRegion#AGS-TF" higherGeopgraphy="Argentina, Tierra del Fuego" country="Argentina" countryCode="AR" stateProvince="Tierra del Fuego" locality={null}
but the following combination would not:
locationID="rs.tdwg.org/ontology/voc/GeographicRegion#AGS-TF" higherGeopgraphy="Argentina, Tierra del Fuego" country="Argentina" countryCode="AR" stateProvince="Tierra del Fuego" locality="seashore on route from city of Ushuaia to Ushuaia airport"
I think this is somewhat impractical.
We have realistic chances of having locationIDs only on a higher geographic level. But these are extremely useful when dealing with geographic uncertainty.
It seems inappropriate to limit the use of gazetteer IDs - where available - for cities of villages to those records not having any further location details. Or do I miss something?
Gregor