- From: Norman Gray <norman@astro.gla.ac.uk>
- Date: Thu, 23 Nov 2006 16:13:57 +0000
- To: semantic-web@w3.org
Greetings, [replying to myself] On 2006 Nov 23 , at 15.54, Norman Gray wrote: >> geo:lat rdfs:subPropertyOf vcard:latitude > > and nothing more. In particular, it would probably be best to > leave the vCard lat/long as informally specified, and let WGS84, > and ITRS, and ..., coordinates to be subProperties of them, for > those few specialists who care. If you want further precision, then the most straightforward way to get it is to give vcard:Location an optional vcard:datum property, whose range is some appropriate URL for the datum in question. If that property is missing, then the error in the location is presumed to be the maximum of 1ulp for the given value, and 10m (which is I think the order of magnitude of the inconsistency between modern datums). What URL? Specs of this sort tend to be paper standards rather than electronic ones, but I'm sure that best-practice representative URLs could be suggested for the important datums. While I'm at it (and now that I look at it), the definition of vcard:latitude and vcard:longitude could be tightened up, by saying that both properties have a range in decimal degrees (as opposed to arcseconds, sexagesimal degrees or radians, all of which are otherwise plausible in some constituencies), and that the vcard:longitude is specifically east longitude (which I think I once worked out WGS84 longitude is, as opposed to west longitude). Also, is the range [-180,180], [0,360], or the corresponding half-open intervals; or is it unconstrained but periodic? There's no need to constrain them to be on earth, though that's another possible appropriate property... All the best, Norman -- ------------------------------------------------------------------------ ---- Norman Gray / http://nxg.me.uk eurovotech.org / University of Leicester, UK
Received on Thursday, 23 November 2006 16:14:16 UTC