RE: [Geopriv] PIDF-LO needs a "derived from another form" method -- an aside

Brian,

Several times now, you have attributed motivations to the use of an
application layer method for acquisition of location information that
are just prejudicial. Please--whether geo or civic or both are provided
is a separable issue from the need for an application layer method of
location acquisition.
 
Any of DHCP, LLDP-MED, or an application layer method, e.g., HELD--
or via manual entry for that matter--would allow for providing one or
both of civic and geo.

Nadine

-----Original Message-----
From: Brian Rosen [mailto:br@brianrosen.net]
Sent: Thursday, July 21, 2005 12:01 PM
To: 'Andrew Newton'; Abbott, Nadine B
Cc: 'GEOPRIV'
Subject: RE: [Geopriv] PIDF-LO needs a "derived from another form"
method

While I don't think it will often occur if location is obtained via
DHCP, it could happen. A more common case I expect is where location is
obtained from the application level protocol being discussed. I think
that some implementations will try to be helpful, and return both a
civic and a geo. One of them will be the "original", and one they will
derive from the original. The argument goes that the users of location
are varied, and some would prefer civic, some would prefer geo, and some
actually need both. An example of the latter is, "find the nearest
Pizza Hut, and deliver a pizza". "Find the nearest" is best done with a
geo. "Deliver" is best done with a civic.

We need to know which data is the original data, and which was derived
from the original data. We have to mark them clearly.

Now, I personally don't think sending both is a good idea, primarily
because I think the recipient will have no idea what the quality of the
conversion is, and we know quality of geocode databases varies
considerably. I think they would prefer to take the original, and
convert it themselves, with a database they understand the quality of.
But that is not a belief shared by all, and if we are going to permit
sending both, we need to mark them appropriately. That's what I'm
asking for.

Brian

-----Original Message-----
From: geopriv-bounces@ietf.org [mailto:geopriv-bounces@ietf.org] On
Behalf Of Andrew Newton
Sent: Thursday, July 21, 2005 11:45 AM
To: Abbott, Nadine B
Cc: GEOPRIV
Subject: Re: [Geopriv] PIDF-LO needs a "derived from another form"
method

Nadine,

This is well stated, and I agree with what you are saying.

On Jul 21, 2005, at 9:42 AM, Abbott, Nadine B wrote:
> May be simpler to just add another parameter to mark
> converted/derived/transformed location.

If I acquire location via DHCP, what I know is that I got it from
DHCP. Not that it was converted or derived.

So I'm not really sure these will be used.

I do agree that if there are two pieces of information, and one is
the original and the other is derived, then they should both be sent
and both marked appropriately as to what they are.

I guess I'm just not understanding Brian's use case of when this
derivation occurs.

-andy

_______________________________________________
Geopriv mailing list
Geopriv@ietf.org
https://www1.ietf.org/mailman/listinfo/geopriv

_______________________________________________
Geopriv mailing list
Geopriv@ietf.org
https://www1.ietf.org/mailman/listinfo/geopriv
Received on Thu, 21 Jul 2005 13:11:42 -0400