Re: Civic Address for V2

Hi Alec,

Thanks for bringing this issue back up.  My $.02:

> a.       I propose we use the same fields as the CivicAddressReport in 
> the Windows 7 Location API. 

Where on earth did you get that idea? :)

I tend to agree with Allan that a representation compatible with RFC 
4119 (more properly, RFC 5139) would be more universally applicable.

I appreciate that some view this as too complex, so there may be value 
in producing a simplified version.  However, it should be a profile of 
the IETF format, in that there should be a 1-1 mapping between the two.


> b.      I propose: Enum {CoordinatesOnly, CivicAddressOnly,  Either}

The IETF HELD protocol [1] has a similar option.  The way HELD solves it 
is to specify a list of acceptable formats, then a boolean "exact" 
option to specify that exactly that set must be provided.  In this case, 
you would define an Enum {coordinates, civicAddress, any} for location 
formats, then have an array of these plus a boolean.  (An empty array 
signifying [any], exact=false)

These two approaches are equivalent when you have the two location 
formats (civic & geo), but the second (HELD) approach scales better when 
there's more than two (e.g., if you allow location to be provided by 
reference).

--Richard



[1] <http://tools.ietf.org/html/draft-ietf-geopriv-http-location-delivery>

Received on Friday, 27 February 2009 18:37:09 UTC