VOCAB: schema.org/box has the description for "polygon"

[dbs]

This appears to have been a problem since at least the 1.0b release (the earliest to which I have access).

[danbri]

Archive.org confirms this has been in error since the start - https://web.archive.org/web/20110925164957/http://schema.org/GeoShape

Nearby but different issue: http://www.w3.org/2011/webschema/track/issues/19
http://geowanking.org/pipermail/geowanking_geowanking.org/2012-June/026179.html

Perhaps something inspired by "A bounding box defines a rectangular region. It is often used to define the extents of a map or define a rough area of interest. " http://www.georss.org/gml.html#gmlbox

e.g. "A bounding box defining a rectangular GeoShape. Can be used to define the extent of a map or an area of interest."

[ManoMarks]

To specify it, we may want to do something Well Known Text-ish. While there's no box spec in WKT, we could do something like 'RECTANGLE(x1,y1 x2,y2)'. Similarly we could have WKT specs for polygon and line 'POLYGON(x1,y1 x2,y2 x3,y3 x4,y4, x5,y5 x1,y1)' and 'LINESTRING(x1,y1 x2,y2 x3,y3 x4,y4, x5,y5)' Circle gets trickier since there's no Circle primitive, but we could do something like 'CIRCLE(x,y m)' with m as meters. We could also drop the primitive names, since we already know the type and just give values.

[akuckartz]

See #113

[danbri]

Mano - thanks. I think the ordering should follow WKT unless there's another compelling model already in widespread use. I'm less keen on using pseudo-WKT syntax as it'll risk leaking out and annoying "real" WKT people. So whitespace-separated comma-separated pairs of pairs seems about right.

[ManoMarks]

Dan, the only issue with that is that there are no circle or box primitives in WKT, which could be an issue.

[danbri]

Aside: W3C is beginning work in this area. http://www.w3.org/2014/spatial/charter

We didn't fix this for sdo-venkman, pushing to next release. Will also liaise with W3C.

[danbri]

For this looming release, I would like to make at least partial progress here:

• currently: schema.org/box "A polygon is the area enclosed by a point-to-point path for which the starting and ending points are the same. A polygon is expressed as a series of four or more space delimited points where the first and final points are identical."
• proposed: "A bounding box defining a rectangular GeoShape. Can be used to define the extent of a map or an area of interest."

Ideal of course would be to have the final sentence giving implementable detail. Can we get that too? (@ManoMarks et al?)

By analogy with polygon, but noting that 2 would be enough, ... "A box is expressed as a series of {x} space delimited points where {y}.".

Any thoughts on x and y? e.g. "2 points", "the points define diagonally opposite corners of a square bounding box."

Any suggestions for x and y?

[danbri]

Is this enough to be an improvement?: "A bounding box defining a rectangular GeoShape. Can be used to define the extent of a map or an area of interest. A box is expressed as a series of 2 space delimited points corresponding to opposite corners of the box."

[ManoMarks]

The most common approach is [minX,minY,maxX,MaxY) where X is longitude, Y is latitude (depending on coordinate reference system). You can see this typified here:
GeoJSON
KML
OSM

[danbri]

Thanks @ManoMarks, I'll try to merge that in. I believe the heritage of what we have here was (possibly GeoRSS-inspired) via IPTC rNews 1.0 (http://www.iptc.org/std/rNews/1.0/specification/rNews_1.0-diagram.pdf) which was added into schema.org very early. http://dev.iptc.org/rNews-1-The-GeoCoordinates-Class ...

Sorry I should have looked there before! For box they do have "A box is the area enclosed by the rectangle formed by two points. The first point is the lower corner, the second point is the upper corner. A box is expressed as two points separated by a space character.". And the other properties do closely match schema.org. Maybe I should ask them about the heritage and use so far of this?

[danbri]

Ok I've marked this one closed for now, although I'm sure other specific Geo topics will re-emerge. It seems that schema.org via rNews is closer to GeoRSS than to GeoJSON.

I also took liberty of adding a little more explanatory text to http://sdo-gozer.appspot.com/GeoShape including some words (inspired by the georss site) retrospectively documenting our ambiguity w.r.t. whether point pairs are comma or whitespace based, i.e. we say we don't care.

[RLRichardson]

Also check out what3words.com where the entire globe has been mapped into 3X3 meter squares and assigned a name consisting of 3 randomly assigned words