RE: on NAT-PT

["Christian Huitema" <huitema@windows.microsoft.com>]

> >An IPv6 host doesn't have to be dead. If it has a piece of logic that
> >combines a "nat-pt prefix" with the IPv4 destination address, then it
> >can send the packet over IPv6 to that fabricated destination, and
rely
> >on the NAT-PT to do the translation job.
> 
> 	it seems to me that modifying an IPv6 host for transition
purposes
> 	is not an option.  basically it makes IPv6 hosts aware of NAT-PT
(or
> 	whatever) translation in the middle, hence similar to RSIP
model.

I don't believe we should have a black-and-white attitude regarding
this. I don't think that hosts developers can ignore that a transition
is going on, at least during the next few years. Clearly, the best way
for a host to support the transition is to be simply dual stack. Do you
want to preclude hosts taking an in-between position, i.e. not
dual-stack but still nat-pt aware? 

-- Christian Huitema