RE: Issue-10 unresolved in meeting today

Ø  I mildly dislike 3 as it is already covered by 2, so redundant.

Disagree. To be able to reference a CRS description with a URI says nothing about how such a reference would be associated with a geometry.

There is a definite lack of consensus here. For example, GeoJSON had a CRS object that applied to the file as a whole [1], though this is now deprecated, probably in favour of a JSON-LD solution [2][3]. Meanwhile, GeoSPARQL [4], though its adoption of WKT and GML, enables (but does not _require_) a CRS to be associated with each geometry, separately. All of these can use URIs, but the pattern for attaching the CRS to the geometry is different.


Ø  4 … is already recorded as separate issue  issue-28,

Good. My intention in making the list was to ensure that the CRS requirements were gathered together. Else there is a risk that the sum-of-the-parts don’t make a whole.

Simon

[1] http://geojson.org/geojson-spec.html#coordinate-reference-system-objects

[2] https://datatracker.ietf.org/doc/draft-butler-geojson/

[3] https://github.com/geojson/geojson-ld/issues/27

[4] http://www.opengeospatial.org/standards/geosparql



From: Kerry Taylor [mailto:Kerry.Taylor@acm.org]
Sent: Saturday, 27 June 2015 9:48 PM
To: SDW WG Public List
Subject: Fwd: Issue-10 unresolved in meeting today




-5 from me.

We have gone round in circles.

I have no objection to 1 and 2, noting that we seem to have lost the http uri part again, which was rather well supported.

I mildly dislike 3 as  it is already covered by 2, so redundant.

I dislike 4 because it puts us back where we started before the last meeting. can we separate the concern of mandatory or not? this was quite controversial when discussed on the email list some time ago.   This is already recorded as separate issue   issue-28, but could certainly be worded better.

Kerry


On 26 Jun 2015, at 10:34 pm, matthew perry <matthew.perry@oracle.com<mailto:matthew.perry@oracle.com>> wrote:

On 6/26/2015 5:06 AM, Andrea Perego wrote:

On Fri, Jun 26, 2015 at 10:06 AM,  <Simon.Cox@csiro.au<mailto:Simon.Cox@csiro.au>> wrote:
Then, the requirement is:
1.
to be able to reference a CRS with a URI, and
2.
to get useful information about the CRS when you dereference that URI.
Then there are at least two more requirements:
3. a mechanism to associate a CRS reference with a geometry description
4. for there to be a default or implied CRS reference where it is not explicit in the data.
+1

Andrea

+1 from me too.

Matt