for a long time the assumption was that IPv6 would be deployed first
in the core, and then in the periphery, of the net. I'm now of the
opinion that IPv6 will be deployed first in the periphery -
both in emerging networks that need large amounts of address space,
and in existing IPv4 nets using 6to4 - and it will be deployed
by folks who have applications that need global address space
(and which perhaps aren't already widely deployed using v4)
and by folks who need to be able to access the new IPv6-only networks.
the emerging networks may be large networks in parts of the world
that are just now getting on the Internet, wireless networks,
and other networks designed to support large-scale data gathering.
(power meters, auto traffic monitors, environmental monitoring,
security systems, etc.)
I think we will have a long period of v4/v6 coexistence, with v4
becoming more and more NATted and popular applications moving
over to v6 based on how poorly they work under NATted IPv4. the
older and better established the application under IPv4, the longer
it will take to move it to v6. SMTP will use IPv4 for a very long
time - not that it won't use IPv6 when available, but for a long time
you'll need to have at least one IPv4-based SMTP server acting as a
mail exchanger for your domain, in order to reliably receive mail.
the core will support v6 when it makes economic sense - i.e. when
top tier ISPs can save enough on bandwidth and support costs (as compared
to tunneling) to make the investment worthwhile. which is not to
say that some major ISPs won't support IPv6 before then.
as for your AM vs. FM analogy - there are a variety of theories about
this, ranging anywhere from artifically making v4 addresses even
more scarce to encouraging a run on v4 address space and making them
scarce that way. but I think the shortage of IPv4 address space
will encourage adoption of IPv6 even without changing allocation policy.
Keith