RE: [idn] case folding

[RJ Atkinson <rja@inet.org>]

At 04:10 12/06/00 , GIM Gyeongseog-KIM Kyongsok wrote:
>   are we going to just use some program (asuuming that 
>somebody programmed based on UTR#21 and made it available
>free of charge) and treat it as a black box?

         My assumption is that normalisation (including case folding)
can technically happen only in one of two places if the DNS
is to remain reliable and interoperable:
         1) DNS client
         2) DNS resolver

         If someone has an alternate proposal that works, I'd like
to have that explained to me. 

>   or is it necessary that most (?) people understand
>UTR#21 ?

         No.  It doesn't work to rely on humans to know some
convention.  Today people can type mixed-case into their URL
window of their browser and case-insensitive results occur.
We need to continue that ability in future.

>   i wonder how many of those who proposed to use UTR#21
>have read and understood it.

         I've read it and think I understand it.

>   I assume that ordinary people will have much diffculty
>reading/understanding UTR#21 (I am not sure if
>ordinary users need to understand #21 though).

         If we do our job correctly, only implementers will
need to understand it.  If we're lucky then a free library
function for normalisation will eventually appear that people
can just use by making a library call to it.

Ran
rja@inet.org