IPv6 seeks to change an existing infrastructure.
no doubt this was the original intent. these days I think it's an
IPv6 seeks to provide services to a much larger number of hosts
than were possible in IPv4, and to provide services (like global
addressing and the ability to support p2p and distributed applications)
which are no longer generally available in IPv4.
this can be done initially without changing the existing infrastructure.
using 6to4, IPv6 is available today for those who have at least one
global IPv4 address. the existing v4 infrastructure will be upgraded
when there is sufficient (anticipated) demand to justify the cost.
In Dave's language, 6to4 makes IPv6 more of an end-point change.
meanwhile, it will be attractive to incorporate IPv6 support into new
the point is that the goal isn't to replace IPv4 with IPv6 so much as
to make IPv6 as widely available as IPv4. once IPv6 is widely
available, it will be attractive to applications developers because
it has fewer limitations than IPv6.
and with 6to4 and most major OS vendors shipping IPv6 (and sometimes 6to4)
in their current product releases, we're closer to having IPv6 generally
available than many think.