Re: [idn] WG Update

["tsenglm@????.??.tw" <tsenglm@cc.ncu.edu.tw>]

> David Hopwood <david.hopwood@zetnet.co.uk> wrote:
>
> > In particular, there have been IDN proposals discussed in this WG that
> > do not support case preservation; Just-send-UTF-8 is one example.
>
> Actually, just-send-UTF-8 supports case preservation just fine; it's
> send-nameprepped-UTF-8 that doesn't.  If the sender already requires the
> receiver to support UTF-8, it might be reasonable to further require the
> receiver to do normalization and case-folding when comparing strings.
> The one exception is DNS queries, where for performance reasons you
> would want to do it at the requestor, but who cares if DNS queries don't
> preserve case (DNS responses still could).
       The reason is not for performance of DNS. We use ACE only for
backward compatability of  existed LDH-DNS server. All LDH-server do
comarison case-insenstively and keep case preserving in response.
       Nameprep will map all Full-Width-english-character in unicode
directly to lower-case ASCII  is considered the confusing of  TLD and  ASCII
name. To force all case to lower case is over-specifications in
requirements. UTF-8 can keep the case of ASCII
but can not preserve the case of non-ASCII . But now AMC-ACE-Z can preserve
one case of non-ASCII , it is an excellent improvements over UTF-8 . We
should consider the specification based on new technology ---AMC-ACE-Z.
      Mapping all Full-Width-english-characters to ASCII is necessary , but
mapping all case of  ASCII to lower case in Nameprep is not necessary .
That will be more compatible with historical  DNS and let all non-ASCII has
the chance to view the name in their like form.
      Only in  case-insensive comparison and case preserving of non-ASCII ,
ACE is more suitable for next step in DNS developing.

L.M.Tseng