There has been a very relevant discussion lately in the geowanking mailing-list "about formal location metadata in blog description area". The point is that since mainstream blog service vendors haven't implemented form input fields for location, it is good enough if we just start sticking raw location information directly _into_ the english text description area of a blog post. . The idea is thus to allow location blogging
For example if you have a blog entry form like this: title : [ Grandma got run over by a reindeer ] link : [ http://www.santaclaus.com ] description: [ Ma grandma jes got run over by one of dem dern reindeers. ]Maybe it is not unreasonable to include location right in the description in some reasonably machine extractable way:
description: [ Ma grandma jes got run over by one of dem dern reindeers { longitude 0 latitude 90 } ]
To be more specific I'd suggest only two formats: 1) Longitude and latitude
{ long:nnn.nnn lat:nnn.nnn }
2) Geocoder or gazetteer with something fairly strict format-wise so the parser doesn't have to be particularily brilliant:
{ street & cross-street, city, state or province, country }
{ place name , city, state, country }
The author (Anselm Hook) also states that:
Why is this not a stupid idea? Because machines will preserve thisinformation... it doesn't require extending the fabric of the web services. The whole goal is to enable writing aggregators that can geo-place posts - in enough volume that it becomes a business case for larger web service providers. (..) Accuracy is not necessarily the goal of this. The goal is just to provide a transport for metadata that is otherwise being stripped or disallowed by naive transmission protocols - and in part to encourage formalization so that in a year or two more blogging services would do it right.