Re: [sipcore] draft-ietf-sipcore-location-conveyance-04

["James M. Polk" <[e-mail]>]

At 11:06 PM 10/26/2010, Thomson, Martin wrote:
>On 2010-10-27 at 12:36:02, Paul Kyzivat wrote:
> > I have never understood why the routing-param is required to be last.
> > And as my examples show, its difficult/impossible to enforce this in
> > ABNF.
>
>If "routing-allowed" is mandatory, it IS possible:
>
>     Geolocation-header = "Geolocation" HCOLON Geolocation-value
>     Geolocation-value  = [ locationValue [ COMMA locationValue ] COMMA ]
>                          routing-param
>
>It's a little strange, but it does work.  Especially if you also add 
>text that states there can be only one GeoLocation header?
>
>(The other thing that I find strange is the fact that there can be 
>zero, one or two locationValue parts.

The pre-Maastricht proposal was to limit to zero or one 
locationValue, but that blew up Richard's rough-loc ID (which defines 
a coarse location for routing, and accurate location for endsystems). 
This is why we made the compromise of allowing a second 
locationValue. If we allow 'many', we need to bloat the doc with all 
that "inserter=" and "inserted-by" crud, and we don't want that, do we?

james

>This is confusing for computer scientists who can only count: zero, 
>one, many.)
>
>--Martin