[geos-devel] WKT EMPTY
artem at pavlenko.uklinux.net
Wed Apr 20 04:23:46 EDT 2005
Martin Davis wrote:
>Can't resist adding my 2c worth...
>To my mind, it's pretty clear that the intent of the syntax spec is that
>words in all UPPERCASE are intended to be literals in the WKT language.
>In this case, the representation of an empty Point is
OK, but then POINTEMPTY is a valid expression as well.
Surely, it wasn't author intention?
>This interpretation also has the advantage of being unambiguous and
>explicit. If the representation was POINT only, there are two
>- an empty Point
>- an incorrect Point representation, with the actual coordinate missing
Well, how about:
Point with coordinate missing == empty POINT
Always treat a 'POINT' as an empty one.
>I guess the spec authors may have the final word, but until they weigh
>in I'd vote for sticking with POINT EMPTY.
Who are the spec authors anyway? Adam Gawne-Cain was involved in
designing WKT/WKB when I used work at cadcorp. Is he the author?
I vote for POINT, LINESTRING, POLYGON etc.
It looks nicer!
More information about the geos-devel