ietf
[Top] [All Lists]

Re: [idn] WG last call summary

2002-03-19 14:00:03
On Mar 19, "D. J. Bernstein" <djb(_at_)cr(_dot_)yp(_dot_)to> wrote:

Paul Robinson writes:
Something *should* be done, but your argument has a hint of
'I never want anything done, ever' about it, which is putting people off.

I have put a huge amount of effort into evaluating the costs of various
IDN proposals. Please read http://cr.yp.to/proto/idnc3.html before you
make any more comments about what I'm trying to do.

I stand by my original statement. Something should be done, but your 
document make you look like a typical whiner - you point out all the 
problems, but offer no solutions to some of the problems you raise. The 
solutions you do offer will take at least 4 years IMHO to be effective, and 
your phase 2 will actually INTRODUCE more problems, as well as allowing the 
continuation of other problems you raise.

You may be aware that I'm the author of the mail software with the
Internet's largest increase in SMTP-server deployment over the past 18
months (reaching #2 in total deployment, behind only Sendmail), and the
DNS software with the Internet's largest increase in domain-name
deployment over the past 18 months (reaching #2 in total deployment,
behind only BIND). My programs relay 8-bit data without trouble, even
though certain people obviously don't understand why this is important.

I know who you are. I know your software. I have spent many hours looking at
your code. It's why I'm suspicious of anything you ever write or say. Don't
get me wrong, I'm not lacking in respect for some of the things you've done,
but I think you're shooting yourself in the foot with the way you'er going
about this. I don't want to have an argument with you, but you're
approaching this from the wrong angle. All very wrong. It's what you do, and
it makes sense to you, but shouting at everybody THAT THEY'RE ALL WRONG AND
IT MUST ALL STOP RIGHT NOW OR ELSE THE WORLD WILL END AND MAIL WILL BREAK is
not the way to get IDNA fixed. Perhaps expanding out IDNC3 to something more
than 'these pieces of software need to be 8-bit clean and then we start
registration' might be somewhere to head next...
 
I want internationalized domain names (and mailbox names and so on) to
work. But I can't support the IDNA proposal; IDNA is a disaster. I also
can't accept having _any_ protocol move forward over so many objections.
Even the strongest desire to _do something_ is less important than the
requirement to obtain consensus for any change.

Don't you understand that to get what you want, all Unix boxen have to be 
*completely* 8-bit OK, all MS boxen have to be *completely* 8-bit OK and a 
good 20% of sites out there will just have to shut down ops permanently? How 
are you proposing to display alpha-ol.com on a VT100?

Had you not considered (and excuse me if I'm being dumb here), that just 
modifying a DNS resolver to understnad PunyCode might just be a nice short 
cut? That perhaps, just maybe, software that recognises e-mail addresses as 
being any continuous text with @ in them, and URLs as beginning http:// is 
going to be able to just throw a string down to a resolver that knows that 
it's doing? And that maybe replacing the DNS resolver on all the machine s 
out there to be able to do lookups with PunyCode might be a TAD more 
realistic than trying to get EVERYTHING, EVERYWHERE to be good with 8-bit?

I must be missing something here... I've *got* to be missing something 
here... please, somebody tell me what it is...

-- 
Paul Robinson