Re: [idn] What's wrong with skwan-utf8?

[Keith Moore <moore@cs.utk.edu>]

> I think some folks find the ACE idea
> intuitively 'dirty', and the UTF-8 idea 'clean'.

I sympathize with folks who would like to return to the world where 
you don't need a separate presentation layer for things like
email messages - you can just copy the octets to a display device
and the right thing will happen.  (not that I think it will ever
happen, but I do sympathize.)  Similarly in the DNS space, if you 
could just copy a domain name to an output device, or accept a 
domain name directly from an input stream, and not have to do 
conversions from/to ACE or any other encoding...that would be nice.
Unfortunately it's just not feasible to retro-fit existing 
applications - at least, not in the near term. 

For new applications this might (to some degree) be reasonable, 
assuming there are any new applications to be written for which 
UTF-8 text is suitable as payload anyway.  And maybe we should 
define a UTF-8 representation of an IDN, in addition to an ACE
representation, in case any new application wants to use it.
But I suspect that in the near term this would cause enough 
confusion that it would do more harm than good.

Keith