Re: [idn] I-D ACTION:draft-ietf-idn-vidn-00.txt

["Brian W. Spolarich" <briansp@walid.com>]

On Sun, 26 Nov 2000, FDU - Sung Jae Shim wrote:

| Of course, in this process of mapping, VIDN can check whether the
| names in ASCII format are active or not on the Internet, and consider
| only those active ones.

  So to resolve name 'foo' I may need to generate O(n) DNS queries?  This
means that VIDN could generate significantly more DNS traffic than current
implementations, which violates requirement #33 of the
'draft-ietf-idn-requirements-03' draft.  It also makes resolver
implementations much more complex than current ones.

  What does it mean to 'check whether the names [...] are active or not'?  

  Section 4.2 of your draft says that 'VIDN includes a coding scheme in
order to make each virtual domain name [...] correspond to exactly one
actual domain name [...]'.  Who assigns this code?  IANA?  If so, your
draft neglects to state any IANA considerations.  How is it generated?  
How does a VIDN-capable application or resolver look up this code?

  You suggest later keeping this code data 'in the mail HTML document at
the server host'.  This approach has several problems:

  - Not all *IDN-aware applications are web browsers;
  - Not all *IDN-aware applications understand HTTP;
  - Not all *IDN-aware applications use TCP.

  I've noticed a tendency among some people discussing IDN to only focus
on the issues particular to web browsers.  While browser-generated traffic
constitutes the vast bulk of Internet activity today, there are a whole
host of other applications and protocols that use domain names in their
implementation, and any proposal that forgets this fact is fundamentally
flawed in my opinion.
  
  -brian