On 1/19/2006 3:17 PM, "John Merrells" <[EMAIL PROTECTED]> wrote:

> 
> On 19-Jan-06, at 8:32 AM, Peter Davis wrote:
> 
>>> In moving identity information between parties it is expected that
>>> the messages of the protocol will include elements that bind property
>>> names and values to digital identities. How a digital identity is
>>> referred to is an important consideration. The properties an
>>> identifier could have are that it allows the user to concurrently
>>> maintain multiple personas, that it could allow for a separation
>>> between the digital identity and the identifier and that it allow for
>>> separation between the identifier and the user¹s agent. In the
>>> interests of flexibility and interoperability we would suggest that
>>> the identifier be a string of characters. This working group may
>>> consider current best practice of what that string might be. For
>>> example, a URL or a UUID.
>> 
>> How about simply that it is in scope to establish a 'uniform
>> addressing
>> mechanism', such as a URI.
> 
> To which piece of the above does that apply? The last three sentences?

Yeah, the identifier part.

> 
> The term 'addressing' worries me a little. That might just be a knee
> jerk reaction from me though, based on my LDAP experience.

But that is what you get whenever you make a string (even opaque) in some
namespace. When I give my dog a name, it's really just virgil.peterdavis
(using, at least, a DNS like delegation notation).

This WG should not invent new identifiers.  Use one of the (too) many we've
got already.  

But I agree wrt the terms use here.  s/address/identifier/;  but we'll get
both in the end.  With the identifier alice, when asserted by foo, you kinda
always end up with an address [EMAIL PROTECTED] (or if you prefer alice!fred ;-)


> A
> key aspect of any information/data model is the separation between
> addressing things and identifying things. In LDAP the DN was
> both, which caused me lots of pain trying to solve the distribution
> and replication problems... which lead to adding the entryUUID
> attribute, so that we had immutable names for entries....
> (I could ramble on for hours about that...)

Yes, I fully agree.  We tend to overload the identifier with other duties.
 
> I'd be happier with 'uniform naming mechanism', which could be a
> URI... people may want something other than an URI.
> 

That works too.

=peterd  (http://public.xdi.org/=peterd)


_______________________________________________
dix mailing list
[email protected]
https://www1.ietf.org/mailman/listinfo/dix

Reply via email to