TCP, SSL, SSH, TLS, HTTP and other tools of mass...

[Allen]

The title on this post is not fair, I agree. The real question I
want to ask is, "Do we ever get it *right* the first time?"

Let's step outside cryptography to look at a possible answer and
avoid the traps inherent in discussing current politics.

 From 1908 to 1927 more than 15 million Model Ts were built.

The Model T got a bunch of things right. Availability, mileage
that we are only now exceeding in a mass way, simplicity. It only
took oatmeal, chewing gum and bailing wire to keep it running.
;-> And it managed all this when roads were hardly more than
muddy ruts.

The use of vanadium steel was one of the key right things as well
  and is part of the reason why many still are running.

And it got a bunch of things wrong, not always on purpose, but
rather that they had not had enough experience to understand
*what* was wrong. My dad told me that learning to drive one was
more than a bit of a challenge given the hodge podge of controls
and they way they interacted. If I recall correctly second gear
and the emergency brake were controlled by the same lever to the
left of the driver.

One of the other things to remember that it cost $850 in 1908 and
about $300 by 1927. Figuring inflation into the price and it
probably only cost $150 or less in 1908 dollars

A lot of this is like early computer interconnects and their
controls. Think back 19 years and what was it like? My memory is
that it was a lot like the early days of cars before the Model T.

All the discussion about the weaknesses and strengths of the
various controls seems a lot like car manufacturing before the
Model T.

So, if history tends to repeat itself, I'd expect a Model T to be
about where we are now, and we should start planning for its
replacement. If we learn from the history of the evolution of the
car we might be able to avoid some of the pitfalls of Edsels,
tailfins and other geegaws. But don't bet on it or hold your breath.

So, back to the core question, did we get the Internet and its
controls right the first time? Nope. Will we get it right the
second time? Nope. Will we ever get it right? Nope.

My feeling is that all the discussions about the limitations or
advantages of what we have right now in terms of controls takes
our eyes off of change, for change *will* happen whether we like
it or not. And not all of it will be for the better as we have
seen with recent events around identity theft, phishing, XSS,
etc. And, of course, the big fish swallowing the littler fish,
making for a mono culture very prone to an unexpected disease.

We might be able to reduce the missteps if we concentrate on
keeping what works and only fiddling with that which is weak.

Hope ever, hope on.

Best,

Allen

---------------------------------------------------------------------
The Cryptography Mailing List
Unsubscribe by sending "unsubscribe cryptography" to majordomo at metzdowd.com