On Feb 6, 2007, at 16:43, chris goad wrote:
Chris, Cameron,
I'm glad to learn that JSON is prominent on OGC radar and already
supported in Geoserver - thnx.
I don't think that "OGC members have suggested writing a white paper
explaining the advantages of JSON over GML, after which JSON might
become incorporated as an alternative format for OGC standards."
equates to "prominent on OGC radar". It sounds more like Cameron and
some other FOSS types got into a discussion at some OGC meeting.
For any alternative to GML to come from OGC would be an incredible
uphill climb, I suspect. (Unless perhaps Google actually manages to
get some semblance of KML approved.) I think any effort to do a Geo
JDIL or JSON would best be initially tackled by an interested group
of people in much the same way GeoRSS was done. Set up a list and do it.
If you really want to pattern it after GML, then consider the GML
Simple Features Profile <http://www.opengeospatial.org/standards/
profile>
Also, are there alternative geo representations out there that might
be a better basis to start from?
Allan
Chris wrote:
At first pass jdil.org feels a bit more heavyweight than I was going
for, doing all the namespaces. And it seems to really just focus on
doing the namespaces
Namespaces are one focus, but JDIL also includes a standard way
of encoding graphs (and not just trees) via a node labeling and
reference mechanism (@id,@ref).
Re Namespaces: In applications where different varieties of data
are being mingled, namespaces are of course the standard
disambiguation technique. Even if a particular job involves
encoding one fixed formalism where name collisions are not a
concern, maybe someone else will be interested in mingling what's
been produced with other content.
I'd prefer to get general consensus on
namespacing json from a standards organization.
Me too, but didn't find any. With OGC moving in this direction ... ?
But past that I'd really like to figure out some consensus on how to
encode lines, polygons, and collections of geometries in to JSON.
Indeed - JSON provides an encoding of trees. JDIL is one way of
extending this to graphs with URI-labeled nodes. Encoding various
domain-specific entities as labeled graphs is the next level in
this direction of work. Needless to say, ports of GML encodings is
one way to go.
-- Chris
----- Original Message ----- From: "Chris Holmes"
<[EMAIL PROTECTED]>
To: <[email protected]>
Sent: Tuesday, February 06, 2007 12:58 PM
Subject: Re: [Geowanking] JSON for GEO
As a side note, I think the Chris Holmes has included JSON as an
output
format for Geoserver.
Yeah, I've been meaning to blog about it and make it available for
download, as a plug in to GeoServer. But got swamped last month.
I drew inspiration from Sean's stuff, but wasn't sure how to handle
geometry types other than points. I was going to call mine 'Simple
Features for JSON', in line with Simple Features for SQL, which
defined
Well Known Text for representing geographic features. It was easiest
for me to do the same - I just used a .toWKT() function to make my
geometry.
But I'm in no way tied to doing things that way, I just wanted to
start
a conversation and hopefully work towards standardization.
At first pass jdil.org feels a bit more heavyweight than I was going
for, doing all the namespaces. And it seems to really just focus on
doing the namespaces. I'd prefer to get general consensus on
namespacing json from a standards organization, I've seen a few
others
talk about it and do it slightly different ways.
If there's consensus on that it'd be easy to just write GML output as
JSON, with namespaces and all.
But past that I'd really like to figure out some consensus on how to
encode lines, polygons, and collections of geometries in to JSON.
Maybe
that just involves me asking Sean to add some more examples on how he
wants to do that in GeoJSON, as I have no strong preference, just
want
to standardize on something.
Chris
chris goad wrote:
Hi,
http://jdil.org describes a simple scheme for implementing
namespaced
vocabularies (including RDF vocabularies) in JSON. The idea is
general, but the examples (and our own applications here at
Platial)
are in the Geo domain.
-- Chris
-------------------------------------------------------------------
-----
_______________________________________________
Geowanking mailing list
[email protected]
http://lists.burri.to/mailman/listinfo/geowanking
--
Chris Holmes
The Open Planning Project
http://topp.openplans.org
_______________________________________________
Geowanking mailing list
[email protected]
http://lists.burri.to/mailman/listinfo/geowanking
--
Allan Doyle
+1.781.433.2695
[EMAIL PROTECTED]
_______________________________________________
Geowanking mailing list
[email protected]
http://lists.burri.to/mailman/listinfo/geowanking
