[GeoJSON] GeoJSON in OpenLayers

[Andrew Turner]

On 3/27/07, Martin Daly <Martin.Daly at cadcorp.com> wrote:
> If GML terminology is the vogue, then geometry is itself a property of
> the feature, and therefore this hierarchy is out of step.
>
> I also worry that trying to pick-and-choose from GML
> terminology/structure might lead to confusion, and that it should be all
> or nothing (within reason, and the "style" of JSON versus XML).  For
> example, "featureCollection" could be used instead of "features".  Of
> course the "all" option is only sensible if compatibility with GML is
> seen as important.

This was what GeoRSS did in using a subset of GML that paralleled
GeoRSS-Simple. So that for every Simple there was a corresponding GML.

However, now this may change b/c as users are pushing the boundaries
of GeoRSS they want features that exist in GML, but aren't in the
GeoRSS-GML. It looks like one solution will be to, in fact, allow
any/all GML in a georss:where element.

So, by choosing all GML, you provide users both with power, but also
possibly confusion/difficulty implementing publishers/parsers.

What about going the route that GeoRSS will be going now which is the
GeoRSS-Simple (point, line, box, polygon) and then GeoRSS-GML (all
GML). This allows for the 'gateway format' of the Simple for people to
quickly put together and use, but then grow into the full GML and
possibly just use existing GML creation/parsing tools.