- From: Andrew Daviel <andrew@andrew.triumf.ca>
- Date: Thu, 15 Apr 1999 03:19:16 -0700 (PDT)
- To: Xavier Plantefeve <XavePlant@iName.Com>
- cc: WWW-HTML List <www-html@w3.org>
On Wed, 14 Apr 1999, Xavier Plantefeve wrote: > ><meta name="geo.position" content="-128.3,48.5"> > ><meta name="geo.placename" content="Vancouver, BC"> > ><meta name="geo.country" content="CA"> > > Isn't there already something defined in the dublin Core for that? Yes, indeed. The Coverage element. However, the gears grind slowly. Recently, the coverage subcommittee has decided to recommend using names for spatial coverage as first choice after a somewhat acrimonious debate, on the basis that around 50% of users are innumerate (is that the equivalent of "illiterate"?) and that incorrect values may be more trouble than none at all. Numeric values are recommended "where appropriate", but the exact form is not finalized. geo.position is partly an attempt to do something useful, and partly to demonstrate that, given enough feedback, users can in fact usually generate correct coordinates. The geo.postion elements are supposed to map 1:1 into dc.coverage.x,y,z if they survive as outlined in the last draft of coverage subelements. The geo.placename element maps more-or-less into the DC coverage.placename element. geo.country uses a controlled vocabulary (ISO 3166) and has no direct equivalent in DC. DC recommends using controlled vocabularies in general but as yet there is no mechanism for allowing an agent to recognize this, while the recommendations for simple use allow more-or-less free text in many elements. If an agent is configured to parse geo.position, it can be easily extended to recognize dc.coverage.x/y Andrew Daviel Vancouver Webpages & TRIUMF
Received on Thursday, 15 April 1999 06:20:20 UTC