Morning Richard,
> >It may be worth keeping in mind some example usages. For instance, there
> >is an existing draft on requesting and disclosing location information
> >within HTTP. In that context, a web browser (perhaps on a cell phone)
> >requests a document and is asked for its location. There is a wide range
> >of possibilities in how it answers that request, depending in many ways on
> >what we end up with here in geopriv.
>
> application specific ...
>
> well please do not go down the road of discussing this in the context of
> HTTP
Yup.
Daviel is attempting to associate some attribute (location) to some HTML
page(s) through a meta TAG.
If one comes at this from an end-2-end or pilc-esque PoV, the network
graph (last mile) is the controlling issue, not the point and associated
probability of sampled correctness in a graph-blind coordinate system.
This in turn becomes more complex when the administrative overlays of
carrier-services-for-mobile-devices (cell phones) is added, and there
the controlling issue is a business issue. Roaming didn't appear to have
been solved prior to the W3C-WAP workshop on mobility and privacy in Dec.
2000 (Munich), and for those thinking about extending the semantics of
P3P (DOM tree element policy evaluation) beyond the wireline demark, the
walled gardens pretty much defined the (in)ability of edge-devices to
either reference rule-sets , or access evaluators, in a location-independent
fashion.
What Daviel could offer, could also IMO be provisioned using CNRP. As there
is a working implementation of CNRP, that does more than just co-location
hinting.
> So what is the privacy object?
Good question. I'm still waiting to find out what an "embedded protocol"
is.
It isn't every day you see an I-D that only references Scott's "WHEN WE SHOUT"
memo. There are a few things clearly left to the imagination, and unless this
is just a General Area policy romp, our good sense requires that we come up
with ... something implementable.
Cheers,
Eric
Received on Wed May 22 12:51:10 2002
This archive was generated by hypermail 2.1.8 : Thu Jan 22 2004 - 12:32:23 EST