[GeoJSON] Point as list of one point, or list of coords

[Allan Doyle]

On Apr 13, 2007, at 17:34, Christopher Schmidt wrote:

> On Fri, Apr 13, 2007 at 04:47:56PM -0400, Allan Doyle wrote:
>> I could get behind different levels of geojson. In fact, it seems to
>> me that if we define it right, people can choose which specific
>> geometry types they support. If you're going to write a "red dot" app
>> that only needs points, why bother implementing all the rest? Better
>> that you feel comfortable using GeoJSON Point objects than feel like
>> you need to invent yet another representation.
>>
>> Of if you're only dealing with GPS tracks, maybe you want Points and
>> LineStrings, etc.
>>
>> GeoJSON - "the Chinese Menu of geo formats."
>
> Testing conformance then becomes a problem of O(numgeoms) rather than
> O(2).
>
> The biggest problem for GeoRSS adoption was lack of validation and  
> clear
> specification. If you don't want to support GeoJSON simple, don't: but
> just don't claim you're conformant to GeoJSON. Explain what you  
> actually
> are.

Having batted this around on IRC, I think what we can do is this

(a) any consumer or producer is free to implement any subset of the spec
(b) any subset that is implemented must follow the spec defining each  
element of that subset
(c) we define at least two "packages" of conformance:

   1. "Simple" - meaning you implement Point, LineString, Polygon, (?)
   2. "Complete" - meaning you implement everything.

(The actual list and names are open for discussion).

I feel quite strongly that we should not require anyone to use more  
than he/she needs. At the same time, I understand the desire to group  
things under an easy to remember name. If the names are just "macros"  
or "shorthand" for groups, we can satisfy both views.

	Allan


-- 
Allan Doyle
+1.781.433.2695
adoyle at eogeo.org