[Geojson] Ellipse spec

[andy e]

No worries about the alt spec - just saw it and was wondering what happened.

Sean, use case for ellipses is pretty basic, I guess. We just need to
process/store them. We do the ellipse->multi-point polygon thing in some
other areas, which is fine, but it's easier to store the original ellipse
details (center point, axii, rotation) and then recalculate that later.
We're currently looking at using MongoDB to store our data and since they
just store JSON (or BSON), we figured sticking with GeoJSON as our format in
the DB was a good idea. I don't think Mongo fully supports GeoJSON (from
what I understand, you just tell it how to index a geo lat/lon field) but
didn't want to reinvent the wheel in terms of what to do with geospatial
json data going in/out of MongoDB.

For right now, we're gonna try something like this:

{ "type": "Feature",
     "geometry": {"type": "Point", "coordinates": [102.0, 0.5]},
     "properties": {"major": "1.0", "minor":"0.5", "rotation":"45"}
}

which probably isn't ideal (we'll have to check all our Points to see if
they have major/minor/rotation properties to determine if they are an
ellipse), but will work for now (just in a prototyping phase).

If you have more questions, lemme know and I can try to answer. But it
basically boils down to: we deal with ellipses a lot and we want to use
GeoJSON. :)

Andy

On Wed, Jul 27, 2011 at 10:52 AM, Sean Gillies <sean.gillies at gmail.com>
wrote:
>
> Richard, would you be willing to take it down or edit it to make it
> clear that it's not version 1.0 to prevent more confusion?
>
> Andy, I'd like to hear more about your use case for circles and
> ellipses. The only one we've discussed here was in the context of
> Richard's geopriv work, and there I felt that fuzzy points were not
> the same as circles. Speaking only for myself, I'm much more open to
> expanding GeoJSON today than I was back when we were just trying to
> ship something small and foolproof.
>
> On Tue, Jul 19, 2011 at 3:23 PM, Richard L. Barnes <rbarnes at bbn.com>
wrote:
> > That's my fault.
> >
> > That document was a straw man proposal for how to integrate RFC 5139
shapes into GeoJSON:
> > <http://tools.ietf.org/html/rfc5139>
> >
> > IIRC, at the time I sent that out, there wasn't much interest in the
more general shapes.  But I would be glad to resurrect the discussion if
there's interest; I still think it's worthwhile!
> >
> > Best,
> > --Richard
> >
> >
> > On Jul 19, 2011, at 5:00 PM, andy e wrote:
> >
> >> I found an early (fake?) draft of GeoJSON that included an ellipse
shape: http://geopriv.dreamhosters.com/geojson/geojson-spec.html (in
addition to circle/sphere/etc)
> >>
> >> Was there any discussion on why this was pulled? Couldn't seem to find
any by searching the list.
> >>
> >> I guess that's not even that important, but we do have a use case for
storing and specifying ellipses. We're using MongoDB to store data and
wanted to store it in GeoJSON, but we obviously don't want to break the
spec, either.
> >>
> >> Our first thought is to just use the properties and a Point to indicate
an ellipse:
> >> i.e.
> >>  { "type": "Feature",
> >>       "geometry": {"type": "Point", "coordinates": [102.0, 0.5]},
> >>       "properties": {"major": "1.0", "minor":"0.5", "rotation":"45"}
> >>       }
> >>
> >>
> >>
> >> Thoughts on that approach? Any thoughts on re-implementing
ellipse/circle back in the spec?
> >>
> >> Is the spec fairly frozen at this point (it appears that way, but I'm
ignoramus as to those sorts of things)
> >>
> >> Thanks!
> >>
> >> Andy
> >>
>
> --
> Sean
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.geojson.org/pipermail/geojson-geojson.org/attachments/20110804/46148460/attachment.htm>