Re: [idn] The layers and character handling in DNS

[Paul Hoffman / IMC <phoffman@imc.org>]

At 9:57 PM +0900 2/15/01, ±Λ_ƒΚΊΖ-GIM_Gyeongseog wrote:
>On Thu, 15 Feb 2001, David C Lawrence wrote:
>
>>  Dan writes:
>>  > If I look at how I would like things to work, if could be like this:
>>  >
>>  > user enters name in application using LOCAL character set.
>>  > ->
>>  > application calls routnes like: gethostbyname
>>  > ->
>>  > gethostbyname converts form LOCAL character set to
>>  > normalised (not namepreped) UTF-8 (or other format).
>>  > it may reject the query if the name contains forbidden characters.
>>  > and sends the query to DNS server.
>>
>>  How does gethostbyname() know for sure what the character set of its
>>  argument is?
>
>i guess that gethostbyname need be extended so that one additional
>param indicates the local char set.

And, with that information, gethostbyname then needs to translate 
completely correctly and authoritatively from that charset to UTF8 
(or at least to ISO/IEC 10646 code points). For this scheme to work, 
we need to have authoritative and non-changing mapping tables for 
every local charset. Even is this was possible (and it's not; see 
below), the software would be huge.

Those of us who have just the easy Latin characters in our names keep 
forgetting that many of the local charsets in the rest of the world 
don't have the mapping tables needed, and won't have them any time 
soon. There is wide dispute about what the mappings should be, and 
even the organizations that have spent the time doing mappings 
*still* are changing them. Part of this is due to the way that 
ISO/IEC 10646 is designed, so it is likely that this issue will not 
be resolved for decades, if ever.

Please, folks: this is a dead end. If you want to keep going down 
this path, maybe preface each message you write with "In an idealized 
universe where there was authoritative and non-changing mapping 
tables for every local charset, ...".

--Paul Hoffman, Director
--Internet Mail Consortium