Re: [Simple] Address and person URIs

[Henning Schulzrinne <hgs at cs.columbia.edu>]

Maybe the right model is to use a scheme, as in

inperson:

Schemes can define any means of communication. Indeed, URIs like mailto: 
(or im:) don't directly define protocols, but rather modes of 
communication. We already have common, albeit non-standard, URIs such as

about:

without any further elements. One could generalize this to

inperson:some-identifier

where the identifier is only locally defined (e.g., it's just the 
presentity URI in our case), with no global semantics.

Jonathan Rosenberg wrote:

> I'm fine to punt, but I think this is a case where the perfect is a 
> serious enemy of the good.
>
> I think that a service URI of the form:
>
> uri:inperson
>
> is sufficient, without any additional per-user data. We can leave it to 
> the human to figure out how to resolve the URI. Indeed, the presentity 
> placetype (home, work, etc.) would probably suffice, since  a watcher 
> can ususally infer how to find them within that place. We just need a 
> way to say, "come talk to me". Does that cover EVERY case? No way. Does 
> it cover many of the important ones? Yes.
>
> -Jonathan R.
>
> Paul Kyzivat wrote:
>
> > Henning Schulzrinne wrote:
> >
> > > > Well, a place isn't always sufficient. We see this all the time in 
> > > > the movies: "Meet me in the bar of tha Acme Hotel at 7pm on April 1, 
> > > > 2005. I'll have a red carnation in my lapel."
> > >
> > >
> > >
> > > I suspect that one of the desirable properties would be that the 
> > > person doesn't change identifiers just because they changed location. 
> > > If you don't do that, it will get very confusing - is this URI the 
> > > same person at a different location or a completely different person? 
> > > (Imagine a friend finder service that works in physical space.)
> > >
> > > There is also the issue that you now have two ways to represent 
> > > "location of a presentity", as a URI or as PIDF-LO, with contact 
> > > information also available as a vCard.
> > >
> > > > So we don't need a *globally* unique person identifier. We need a 
> > > > person identifier that is locally unique in the context of the time 
> > > > and location of the intended meeting. And the time is important too.
> > > >
> > > > The time we can deal with. With regular status it is NOW. We can use 
> > > > timed-status to cover some other time.
> > > >
> > > > I don't think location has to be part of the Contact, as long as it 
> > > > is provided as part of the (timed-)status along with the contact.
> > > >
> > > > Coming up with URIs for identifying people locally within a 
> > > > particular time-space region sounds like future research. Sounds 
> > > > like some sort of arbitary predicate would do the trick.
> > >
> > >
> > >
> > > Do you want to define people by a game of twenty questions?
> >
> >
> >
> > Well, it would be easier if everyone just agreed to have a barcode 
> > branded on their forehead and an RFID chip embedded in them, each 
> > containing their UN assigned global person UUID.
> >
> > But I sense a lot of people don't want to be so easily or irrevokably 
> > identifiable. So weak forms of identification may be all that is 
> > generally acceptable. (Hence "I'll have a red carnation in my lapel" 
> > rather than "The SSN on my forehead will be 123-56-7890.")
> >
> > Unfortunately, this doesn't lend itself well to URIs or URNs, because 
> > the information isn't Unique.
> >
> > I think we continue to punt this one for now.
> >
> >     Paul

_______________________________________________
Simple mailing list
Simple at ietf.org
https://www1.ietf.org/mailman/listinfo/simple