[uf-discuss] geo - accuracy of coordinates

[Kevin Marks]

Andy, you're missing the point.

A bare lat-long pair is not always helpful.

If that's all you have, you can't really display a useful map. The 
existing mapping tools tend to use product-specific ways of specifying 
the degree of zoom needed, to distinguish between the right side of my 
desk, the Technorati offices, South of Market, San Francisco, The Bay 
Area, California, and America.

A radius of interest is a way to express this in a platform-neutral way 
that doesn't require address-parsing. It is readily derivable from any 
specific mapping platform.

If you want to express a geographic feature such as those I mentioned 
above, clearly a human-readable label such as 'San Francisco' is a 
better than a polygon. You probably realise that the polygons necessary 
are in fact fractals, with the resolution necessary determined by the 
zoom level too.

We don't want to replicate Arc-Info here, we want to replicate useful 
user-info.

On Oct 2, 2006, at 2:58 PM, Andy Mabbett wrote:

> In message <A919CF31-0EDC-4BB2-85B5-8BEC6CC7011F at lava.net>, Colin
> Barrett <timber at lava.net> writes
>
>>> Or the capacity to describe a polygon...
>>
>> I call the 80/20 rule into effect here.
>
> Fine, I'm confident that more than 80% of countries, counties, towns,
> cities, gardens, parks, nature reserves, and industrial estates are
> polygons, and fewer than 20% are circles.