Re: [idn] iDNS re-chartering proposal, take 2

["Eric A. Hall" <ehall@ehsco.com>]

Dave Crocker wrote:

> >  Going with ACE-only directly equates to two mandatory encodings.
> > This is an incredibly high cost to lay on the Internet
> 
> Thank you for that economic assessment.  As has been pointed out,
> however, it is not valid.

This is an outrageous assertion. It is not possible for two encodings to
be as efficient for the community as a single shared encoding. Using the
same encoding for DNS as is used for the protocol, the document formats,
the application I/O and the underlying operating system is the maximum
possible efficiency for the Internet commmunity as a whole. Conversely,
introducing an entirely separate mandatory encoding that has no tools, no
history, no experience and no users, is the antithesis of operational
efficiency. Claiming as much is unbelievable. I'm astounded by this.

> However, it is nice to see that you agree that UTF-8 is also an
> encoding.  As such, the relative difference in cost, between an ACE and
> UTF-8, is negligible.

Negligible? Does the phrase "you are right, ACE costs more to implement"
stick in your throat or something?

I have said several times quite clearly that ACE is necessary as a
backwards compatibility mode for legacy applications. Where we part is
that you also believe that this second mandatory untested encoding that is
jammed into a transliteration layer for every existing and future
application is the ultimate in practical design. Just admit that it has
problems which can be addressed by an optimized form with a different set
of priorities. Come on, just say it!

-- 
Eric A. Hall                                        http://www.ehsco.com/
Internet Core Protocols          http://www.oreilly.com/catalog/coreprot/