[Geowanking] RE: New coding schemes - geospatial metadata

[stephen white]

On 18/06/2007, at 11:56 PM, brian grant wrote:
> correct but a hierarchical scheme originating from a recursive  
> function can
> be very database friendly, open and easy to understand.

I don't really agree... the recursive function defines a fixed shape  
like triangles or octagons, and you can say that your data point is  
somewhere inside this shape which is about that size by the depth in  
the recursion.

You couldn't use it to define an irregular smear or vague areas  
defined more or less at random. Think of pizza delivery areas where  
the range of coverage is defined by the roads in the area rather than  
a circle around the shop.

Bezier curves are more useful for that kind of outlining, but also  
suffers from hard-edging ... it's sometimes more useful to define  
shapes by gradient rather than contours, so even algorithms like  
OpenGL lighting with surface normals can be more appropriate.

With such a wide range of marking out areas, including gas density,  
time-lapse for dissipation, etc, why would you define your new world  
as "triangles all the way down"?

--
   steve at adam.com.au