Re: [Geojson] Question regarding RFC 7946: Interior and Exterior Linear Rings

[Sean Gillies <[e-mail]>]

Hi Chris,

On Tue, May 23, 2017 at 9:19 AM, Chris Goodman <
[e-mail]> wrote:

> Polygons with more than one of these rings, the first MUST be
>
>       the exterior ring, and any others MUST be interior rings.  The
>
>       exterior ring bounds the surface, and the interior rings (if
>
>       present) bound holes within the surface.
>
>
>
>
>
> Q1. If an interior linear ring (hole) is not completely inside the
> exterior ring (border) of a polygon, is it an illegal polygon?
>

>
> GeoJson.io allows such an interior ring to be imported. Perhaps it is a
> bug.
>

It would be invalid from an OGC simple features perspective but is not
invalid GeoJSON.


>
>
> Q2. If the interior ring is not clockwise, then what is the expected
> behaviour of the polygon?
>

The same.


>
>
> Q3. Should ‘interior linear ring’ and ‘exterior linear ring’ be formally
> defined in the RFC?
>

The WG didn't find this necessary. Are you saying it would improve the
spec? How so?


>
>
> Q4. If two interior linear rings overlap, then is the result the union of
> the two holes?
>

Geometric operations are outside the scope of the GeoJSON. The format only
describes a serialization for polygons. Exterior ring comes first, interior
rings follow.

>From a OGC simple features perspective a polygon with overlapping holes is
invalid, but I've found that GIS software often has fault-tolerant modes
that can deal with these kinds of invalid shapes. PostGIS's buffer(0.0),
for example, which I expect would indeed replace the overlapping holes with
a single hole that is their union, and not create any "anti-holes" or new
positive areas.

-- 
Sean Gillies