>>>>> Gavin Henry <ghenry...> writes: >>>>> Ivan Shmakov wrote: >>>>> Kurt Zeilenga <Kurt.Zeilenga...> writes:
>>> As an alternative approach, I would suggest having a single >>> multi-valued attribute that would contain URIs expressing the >>> location of the attribute. This would push various issues, such as >>> which location system is being used, out to the URIs. This is a >>> good thing as such issues are not unique to LDAP/X.500. >> I've never heard of the geographic coordinates being encoded as an >> URI. Could you please give any references on that matter (if >> there're any)? > There are examples in the draft the Kurt listed above. FYI: > http://www.ietf.org/id/draft-ietf-geopriv-geo-uri-01.txt It was noted (e. g., the thread started with [1]) that the goals of this scheme are simplicity, interoperability and human-readability, hence the choice of a single coordinate system (WGS 84) in wide use today. However, for the purposes of passing location references in a mixed-coordinate system environment (and the geospatial applications use a sheer variety of these), the only coordinate system is not enough. Also, while the newer revisions of the draft are expected to allow for the spatial uncertainty to be specified, this specification is to be made in meters, which also has some implications (consider, e. g., [2]; well, to put it simple, consider that a temperature in Celsius degrees is given an uncertainty value in Fahrenheit; and when it comes to the spatial coordinates the issue is even worse.) Therefore, while the geo: URI may be taken as the basis, its use (in its current form) for geospatial metadata (which is my primary interest) doesn't seem feasible. [1] http://www.ietf.org/mail-archive/web/geopriv/current/msg07711.html [2] http://www.ietf.org/mail-archive/web/geopriv/current/msg07695.html -- FSF associate member #7257