Rather then "invent" a specific approach for targetArea, why not follow ISO
19107 as the abstract model and GML as the "implementation" instance? Polygon is
defined in the OASIS "where" draft and in the requirements document (which is
available in the Kavi document archive). Very clearly stated. We should be
consistent with how ISO, OMA, IETF, OGC, and others are defining polygon.
Cheers
Carl
----- Original Message -----
From:
McGarry, Donald
P.
To:
Cc: Carl Reed
()
Sent: Monday, May 03, 2010 9:41 AM
Subject: Polygon in TargetAreaType
Carl and others-
Question:
The Polygon element in targetArea is a list of lat lon
pairs that represents a polygon. However, one problem that I have noted
when storing polygons in a geospatial database, or exposing to GIS tools is
RingOrientation. Notably, you normally follow the “left-hand” rule,
which states that if you are walking the polygon you left hand should be
pointing towards the center of the polygon area, which results in a
counter-clockwise representation of points in the polygon. If this is
the case, I think our example in the DE documentation was constructed using
the opposite linear ring orientation. When validating a polygon, do we
assume a certain ring orientation, or state that it can be either or, and just
take the case that doesn’t span multiple hemispheres?
Don McGarry
The MITRE Corp.
Office: 315-838-2669
Cell: 315-383-1197
