Re: Federation protocols

Dnia sobota, 1 czerwca 2013 o 04:01:18 Miles Fidelman napisał(a):
> Melvin Carvalho wrote:
> > On 1 June 2013 01:10, Mike Macgirvin <mike@macgirvin.com
> > 
> > <mailto:mike@macgirvin.com>> wrote:
> >     I pointed this out earlier but it got lost in the interim
> >     discussion - from a Red point of view, any DNS-based name is
> >     transient. So we cannot easily inter-operate in your DNS-based
> >     world. I am "Mike Macgirvin".  At the moment I might be located at
> >     mike@zothub.com <mailto:mike@zothub.com> - tomorrow I might post
> >     from george@jetson.com <mailto:george@jetson.com>; and still be
> >     seen to my friends as Mike Macgirvin. If you
> >     subscribe/follow/whatever either of these webfinger ids from a
> >     traditional "federated social network", you'll miss many of my
> >     posts, and I won't see many of yours. They're going to or from a
> >     different DNS-based location. We didn't do this to be different,
> >     we did this because of a clear need in our communities for such
> >     mobility.
> > 
> > The limitations of DNS are apparent, however it does have advantages
> > too.  Not least that it has a massive network.  It seems problematic
> > to build a social network that is not based on DNS.  It is a common
> > strategy to try and reinvent DNS, tho no one has done it yet, and I
> > suspect no one will have done it in 5 years time either. However, I
> > may have completely misunderstood your point :)
> > 
> > As an aside, we still dont really know what webfinger is going to be,
> > it has not yet become an IETF standard and I think it's felt there are
> > some critical shortcoming, which may or may not get fixed shortly,
> > time will tell
> > 
> >     Some will respond that WebID is the obvious solution - not really.
> >     I don't want to carry an identity dongle with me when I'm at the
> >     university in the computer lab.
> > 
> > You may want to update your understanding of WebID. WebID has evolved
> > to just be about using HTTP identifiers as profiles, similar to
> > tent.io <http://tent.io>.  Dongles and other authentication methods
> > are orthogonal.
> 
> You know there are URI (and other) schemes that DON'T involve DNS or
> hostnames.  A few that come to mind:
> pretty much all the URN schemes
> X.500 & LDAP directories
> CNRP (Common Name Resolution Protocol) - http://tools.ietf.org/html/rfc3367
> the "tag" UI scheme RFC4151
> digital object identifiers (DOIs)
> UUIDs

Exactly. I think using URI (with an optional "username@" part) as UID makes 
sense and doe snot tie us to DNS. Think of the TOR network - nothing is 
stopping anybody from using 'user@example.onion' as an UID, and that is 
*completely* outside the DNS hierarchy.

The "shape" of the UID doesn't mean it is anchored in the current DNS system.

> Of course they all beg the question of how to maintain a namespace, and
> then maintaining resolution infrastructure, and all those nasty issues
> of distinguishing among multiple people with the same name.

Indeed.

-- 
Pozdrawiam
Michał "rysiek" Woźniak

Fundacja Wolnego i Otwartego Oprogramowania