To all:
draft-liana-idn-step is proposing Romanization
for all languages in DNS only, as a script block specific ACE.
Advantages are: 1
to reduce fonts of different scripts to confuse users.
2. It also proposes for [nameprep] to include all other local
code standards, such that where an IDN is supported,
it doesn't matter what display code it is, a user always
finds a local input keyboard to be used and the hostname
is properly [nameprep]ed to be convered to access DNS.
[nameprep] is huge already. To add a few corresponding
subsets codepoints will not increase too much of its complexity.
3. I suspect it may speed up [nameprep] processing, since it
limits many scripts to very specific codeblocks quickly, for
example Kana script, to very small sections of [nameprep].
4. In fact, for each hostname, there will be more than one
presentations to its users, an ASCII version and a local script
version. For Chinese, I can see there may be four presentations:
one for the Mainland users, one for Taiwan users, one for Hong
Kong users, one in ASCII. (Traditional and simplied characters are
often used together in Hong Kong.)
Liana
