As Martin points out in his return to blogging (welcome back !!) one of the most interesting presentations at this OGC TC meeting was the one given by Raj Singh on GeoRSS.
GeoRSS is a rapidly developing standard to encode locations into RSS feeds, so that for example, your blog aggregator software could subscribe to blogs whose entries relate to locations within 50 km’s of where you live, and of course to allow this to happen, also allow authors to tag their content with geographic location.
The real challenge here is to keep this encoding simple, following the rationale of RSS itself, while also allowing for more complex geographic information than simple Lat,Long pairs.
The approach taken so far is too offer two encoding types simple and gml which work with the Atom standard to extend RSS.
The Simple encoding which will meet the needs of most users take the form;
While the more flexible GML encoding embeds GML features, offering more flexibility including different co-ordinate systems, and uses the form
<gml:Point> <gml:pos>45.256 -71.92</gml:pos> </gml:Point>
Both encodings also support tagging line-strings, polygons and envelopes (MBRs).
There is real potential here for a standard to bring geographic information to the mainstream, however we also need to recognise there are other encoding out there and geoRSS will have to make it’s case against gpx, kml etc.
Look out for geoRSS entries from edparsons.com in the near future !
Written and submitted from The Marriott Hotel, Huntsville, using the hotel in-room internet connection.
2 replies on “GeoRSS – bringing geography to the blogs..”
As someone interested in the overlap between the web and geospatial technologies, it’s great that standards are being/have been introduced and adopted.
However I do have issues with not explicitly specifying which coordinate is the latitude and which is the longitude, leaving it up to the user to get it right. From an end-user point of view, I think it makes more sense the W3C geo (point) way, where you state geo:lat and geo:long as separate entities. Admittedly it is a little more verbose, but it helps the user get it right.
As an example of coordinate data that’s not been classified, take a look at GeoURL. You can clearly see the points where users have accidentally flipped coordinates because, like the GeoRSS standard, they are not required to say which is lat and which is long.
I agree there is a problem, however coordinate reference systems themselves, like WGS84 used by GeoRSS as default DO specify the order to be Latitude, Longtitude – which I would argue is also the wider convention.