Issues lists and the "preprocessing" topic

[JFC Morfin]

Mark, Patrick,
for years I advocate in vain this to be clearly defined. Either IETF 
has an open mind about mapping and designs mapping independent 
protocols or it decides to delegate mapping to Unicode and organise 
it like when delegating Names and Numbers to ICANN. End to end 
protocols should be user application mapping independent. In the no 
end to end IDNA case there is a problem as a pseudo-presentation 
layer is introduced at user application layer, more over some ccTLDs, 
such as ".su" (now confirmed by ISO 3166) and several other ccTLDs 
have a mapping independent registration policy.

However, there is no ambiguity if the WG-IDNABIS Charter is 
respected. As answered by the Chair, the target is not a universal 
multilingual DNS support, but an IETF IDNA2003 consistent evolution 
in order to make IDNA work better. IDNA includes mapping. Patrick's 
position is the position this WG is to adopt. Mark's position is an 
interesting alternative that other non-IDNA solutions could seriously 
consider, moreover if Unicode can come with a solution to restore the 
information lost in mapping.

This raises an issue I hoped this WG could be spared which is vital 
for ICANN, WIPO and TLD Managers : the same ISO 10646 label can be 
registered to be used with different mapping. Depending on the 
algorithm being used the same ML domain name presentation will be 
supported by different ASCII domain names and resolve differently. 
This results from the basic lack of analysis between DNS presentation 
and application layers. I was opposed on this because they look the 
same in ASCII and people forgot that RFCs are not exclusive standard.

IMHO, the only way to address this is in speeding up the WG-IDNABIS 
process, introducing it as the way to form multilingual domain names 
for current ASCII English users, with the hope that the usage wave 
will call for interoperable alternatives and delay the problem until 
the community has found a solution.

jfc


At 05:47 26/08/2008, Patrik Fältström wrote:
>On 21 aug 2008, at 01.59, Mark Davis wrote:
>
> > Based on comments from John and others at the meeting, however, it
> > appears
> > that the working group is fundamentally not interested in having a
> > common
> > specification for a mapping phase be part of the IDNAbis, and that
> > it would
> > be better done by organizations like Unicode or others. Based on
> > that, I
> > modified the draft at http://docs.google.com/Doc?id=dfqr8rd5_51c3nrskcx
> > , and
> > submitted it to the UTC for consideration.
>
>A few things:
>I have not seen this consensus you refer to. The contrary. I see
>support for such a document.
>
>I personally support such a document in the IETF, and have told you
>and everyone I have been asked.
>
>The URI you have above does not resolve. I get back a 404.
>
>Regarding earlier versions of this document, I think it is a fine
>start, but too bound to "web based applications". I.e. ok as such, but
>if applicable to more protocols than web/http, it could be more
>generic. But see comments on in what order the wg have cycles for
>working on multiple documents and multiple threads.
>Patrik