Tony, 

First, there are a different uses for identifiers and each has
different requirements.  Before we can talk about properties for
"identifiers" in general we need to separate out which particular uses
we are talking about.  For example, are you including discovery
starting from FQDNs?  Kerberos identifiers for auth/auth?  

Second, what is the purpose of the exercise?  I suspect that every
practical use case knows what properties it needs already for its
various identifiers.  If you look at the requirements of a specific
use case the properties fall out.  One possible purpose I could see
(and one I was trying to get at in my slides) would be to reduce the
number of identifiers we think we need.  What do you have in mind?

Scott

Excerpts from Tony Li on Tue, Apr 14, 2009 08:52:00PM -0700:
>
> Hi all,
>
> Thank you for your help with the terminology discussion.  I'd like to  
> propose the next question: what are the necessary and sufficient  
> properties of identifiers?
>
> We discussed this at Dagstuhl and came away with the following:
>
> - topology‐independent
> - stable, at discretion of owner
> - unambiguous
> - distinguishable (two for same object?)
> - facilitate session‐level referrals
>
> Thoughts?
>
> Tony
>
> _______________________________________________
> rrg mailing list
> [email protected]
> http://www.irtf.org/mailman/listinfo/rrg
_______________________________________________
rrg mailing list
[email protected]
http://www.irtf.org/mailman/listinfo/rrg

Reply via email to