Peter,
It does often seem to be the case that poorly designed short-term
solutions to problems are adopted before well-designed things that
work well in the long-term, particularly when the long-term solutions
come with a greater transition cost. NATs made more sense than IPv6
for a certain subset of popular applications meeting certain criteria;
the error was in people assuming (or being misled) that these were
the only applications of interest. Now people are proposing solutions
to the NAT problems which are more complex than IPv6, at a time when
IPv6 is becoming available "off the shelf".
The "fight" isn't "over" because the Internet continues to grow, and to
evolve quite rapidly and probably will continue to do so for quite some
time. Neither NAT nor IPv6 (as we know it now) will be the terminal state.
And while I would be the first to admit that the entire Internet
suite of protocols didn't spring fully formed from Athena's head, neither
do I buy an argument that assumes that everything worthwhile occurs
organically and that natural selection in the marketplace is the
only force that matters. A great deal of the success of the Internet
is due to some good solid design in IP and TCP. These did not crop
up at random, and they did not come from a private vendor; and they
proved superior to competing technologies from both vendors and ISO.
And if for example OSI had won instead of TCP it is difficult to imagine
how the web would have succeeded under such conditions.
In the physical world, bridges are designed, not discovered. It
requires substantial investment and usually inconvenience to build
them; they don't just happen by accident. But when a bridge gets too
weak or the traffic load gets too large for it, we don't argue that
the bridge is as Nature intended. Instead, we build a better bridge.
Keith