domain of lat/long too constrained; use spacialThing, please

I want to use lat/long to describe
airports, meeting rooms etc.; they're
not points, are they?

"If we encounter a description of a Point, call it ?X, and another, call
it ?Y, and ?X and ?Y have identical values for their 'lat' and 'long'
and 'alt' properties, we can conclude that ?X == ?Y, ie. that ?X and ?Y
represent the self-same thing."
  -- http://www.w3.org/2003/01/geo/

That seems too constrained.

Suppose I'm in the KC airport... we might say

  :DanC geo:lat "39.2975"; geo:long "-94.71389"; geo:alt "0".
  :KCIAirport  geo:lat "39.2975"; geo:long "-94.71389"; geo:alt "0".

Do you really want to conclude that

	:DanC = :KCIAirport.
?

You might say that my lat/long isn't *exactly* the
same as the airport's, but it's a real hassle to
make sure no two things ever get the same
lat/long, right?

Perhaps I should say
	:DanC geo:atPoint [ geo:lat "39.2975"; ...].
but that's a tedious and, imo, useless indirection.


I like the design of...
  http://www.cyc.com/cycdoc/vocab/spatial-vocab.html#SpatialThing

hmm... Austin and Texas are considered points in...
  http://www.cyc.com/cycdoc/vocab/spatial-vocab.html#pointOnPath

but note that
(per http://opencyc.sourceforge.net/daml/cyc.daml)
the domain of pointOnPath is cyc:Thing, i.e. 
rdfs:Resource.

Looking at the test data...

<geo:Point>
<rdfs:label>abingdon</rdfs:label>
<geo:lat>51.4017594</geo:lat>
<geo:long>1.1757167</geo:long>
</geo:Point>

abingdon is a point? That doesn't appeal to me.


some background questions...

"The origin of this workspace was the 2003-01-09 discussion in the RDF
Interest Group IRC channel."
 -- http://ilrt.org/discovery/chatlogs/rdfig/2003-01-09

Oh... yeah... I remember that. I don't recall anybody agreeing
to anything in particular, but no matter...


  "Currently we have only a very minimalistic RDF vocabulary"
	-- http://www.w3.org/2003/01/geo/

who is "we"? The page is only signed by you. It claims
"we" are accountable to the #rdfig channel... hmm...
not a very reliable mechanism. If I want to participate,
do I need to read all the logs? I'm using email,
since you're not in #rdfig just now, with copy to
www-archive+geo (unprecedented/unreserved, I believe).


-- 
Dan Connolly, W3C http://www.w3.org/People/Connolly/

Received on Thursday, 23 January 2003 02:35:38 UTC