Re: [Geopriv] comments on draft-linsner-geopriv-relativeloc-02

[Marc Linsner <mlinsner at cisco.com>]

James,

> 
> 1) It is still unclear why this need to implemented as fields in a civic
> location object when it has been pointed out repeatedly that the
> relative component is:
> a) applicable to geodetic coordinates as much as it is a civic
> location
> b) and offset of, not an integral part of the base location

The users of the civic relative location information believe it is part of
the civic location as the resulting location is within the boundary defined
by the civic address.  I attempted to make that point clearer in this new
version of the draft.

Your suggestion of designing a compound location may in fact be the best
approach when the resultant location is not contained with the boundary
describing the reference point (the civic location in this case).

> 
> The concept is therefore best implemented as a compound location
> consisting of a reference and an extent. Both Hannes and I have shown
> how this can be done in a sensible fashion that does not involve
> conflating the existing civic form. Without a very sound argument I can
> see no reason to blend the civic and relative forms into a single object
> that precludes the use of a geodetic reference point.

Can you offer technical reasons why extending the civic location document is
bad?  (As in breaking a protocol/privacy/security)

Nothing precludes anyone from adding relative location to geo, or creating
the new relative location type you are suggesting.  It's simply that in this
context, having the information contained within the civic document best
fits the use case, per the users.

> 
> 2) You have a field RESLPOS-UOW I am guessing that this is supposed to
> be UOM. I would strongly recommend that use the measurement URIs defined
> by the OGC, rather than an enumerated token.

Will fix in the next version.

> 
> 
> 3) Section 5 is most confusing. You can't just extend the schema in the
> fashion that you are trying to do. You need to create a new namespace,
> and you can do this in one of two ways.
> a) You create a new namespace just for the refer elements, or
> b) You create a new namespace, extending the existing schema and
> include your new elements.
> Either way it has to have a new namespace and you will need an
> example showing show how to use it.

Will fix in a future version.

-Marc-



_______________________________________________
Geopriv mailing list
Geopriv at ietf.org
https://www.ietf.org/mailman/listinfo/geopriv