Wikidna team final report

Andrew Sullivan ajs at
Thu Sep 24 16:28:27 CEST 2009

On Thu, Sep 24, 2009 at 03:52:20PM +0200, Dominique Lacroix wrote:

> And overall, for 25 years, users ARE USED to perpetual changes in the  
> Internet, and need to always learn and train.

If you are suggesting that DNS names have changed dramatically in that
time, then I defy you to provide the examples.  There are two obvious
ones.  The first was the alteration of domain names to permit digits
at the beginning of labels, so long as the top level label did not
start with a digit.  The second was the addition of IDNA and the
consequent possibility to use non-ASCII labels.  

But both of these were _additions to_ the available labels, and not a
fundamental alteration to already-working syntax (which continued to
work just fine).  Also, the second case has still not settled: if it
had been completely satisfactory to everyone, we wouldn't have needed
to undertake this work. 

What you propose is to break an already-established rule, and alter
the way that works.  In terms of API, interface, or any other design,
that's always a terrible mistake.  Automobiles still have steering
wheels, even though we've had the technology for decades to control
cars by some other means and even though the steering wheel was for a
very long time one of the primary sources of injury in accidents.  The
fact is that retraining all the drivers in respect of such a
fundamental change in the automotive interface is not really an
option.  It's too great a violation of expectations.  The same thing
goes with respect to case-insensitivity of DNS names.

> You seem to say: "Domains names are not names, domain names use words  
> but they are not words."
> Please allow me to feel a little doubt.

You need to read the output of this WG, then.  Domain names are names,
but they're not names in any language except "the Internet".  They
"use words" in the same sense that companies use words when naming
their products (tell me: was "Pepsi" a word before the cola company
used it?  What about "Viagra" before the little blue pill was
invented?)  They are not words in any language, they never were, and
they've never obeyed the spelling rules of any natural language on
Earth.  This point has been made over and over again on this list and
in the WG documents.  Even if you wish things were otherwise, they're
not, and I urge you to stop going on about this.  Your premise is
false, and therefore your conclusion is not proven.



Andrew Sullivan
ajs at
Shinkuro, Inc.

More information about the Idna-update mailing list