Re: Notes from Non-meeting

> This is true: we do not have unique identifier for people. But that doesn’t
> hinder you to ask where “Jorge” is, although many call me
> "Ricardo". If the problem is that the name
> “Jorge” is not unique, well you could replace it by “Jorge.Cuellar@mchp.siemens.de” or whatever.

Certainly, but we're not designing a generic directory protocol here.

> The fact is that people have names by which they can be
> identified in certain contexts. For instance the
> Location Server may distribute names to the people
> who want to use the service (both for client and for target roles).
> More importantly, there are many types of credentials that you
> may obtain to prove ownership of your identifier.
>
> Ipsec, PKIX and many other groups at the IETF do not have problem with this.

I didn't say you had to have *exactly one* identifier, but at least one.

>
> Your argument applies not only to people, but as you note, to buses,
> Taxis and other objects which do not have a single ID. For instance laptops,
> handhelds, etc. Should the usability of geopriv be restricted to only cell phones, with a unique identifier? (Not all cell phones will have this
> property either).

Again, the identifier doesn't have to be unique in the sense that one
object can have multiple identifiers. But it gets hard if one identifier
refers to multiple objects that may not always be in the same physical
place. Also, you need at least one - location information without some
protocol-representable idea of what it is for is pretty useless. This is
another argument why this problem gets much easier if you embed location
information in existing protocols since you usually already have means
to identify objects of interest.
Received on Thu Mar 21 17:53:21 2002