Re: ISSUE-19 (point-encoding): How should we represent points? [Core FPWD]

comments inline...

On Jun 6, at 1:59 PM, Thomas Wrobel wrote:

> Really not sure about merely having space-separate lat/long/alt co-ordinates.
> This means we arnt specifying a name field for them - how well does
> this work with none xml formatting like JSON? (co-ordinates will need
> to be passed into separate, likely double, variables for use after
> all).

Specifying lat/long/alt in attributes adds clarity if you have a single point, but when you start dealing with lines and polygons and you have -- as is often the case with natural features like shorelines -- thousands of points along the line, that clarity bloats the message. That's why the practice has evolved in the geospatial community to go with a format that's as terse as possible when it comes to the coordinates.

> Also, does the point itself need an ID ? (the POI itself is required
> to have a unique one, but does each point it might use within it
> also?)

No the point doesn't need an ID. Most elements in GML can have an ID, and I copied that from an example where it made more sense for <Point> to have an ID. Please ignore that part of the verbose example.

Received on Tuesday, 7 June 2011 00:12:37 UTC