local mappings

John C Klensin klensin at jck.com
Sat Jan 24 06:11:09 CET 2009



--On Saturday, January 24, 2009 5:46 +0100 Patrik Fältström
<patrik at frobbit.se> wrote:

> Quick reaction before coffee:
> 
> First, As I have said many times, I support specifications of
> mapping rules.
> 
> But, I do really see a need between what is mapped to
> something and what is possible to use in DNS.
> 
> Because of this, I do not as an immediate reaction like having
> both uppercase A and lowercase having the same property,
> PVALID.
> 
> Create a new property, and divide rule Stable in two, and I
> think I am with you.

Quick reaction (late at night and post-drinking):  This possible
mapping change isn't free.  It costs us the complete,
no-information-loss, symmetry between the Unicode and ACE forms,
which I still believe is an important consideration.  I can't
even begin to think about terminology right now, but it may help
to make a distinction between a map-able label and a label that
could be obtained by converting an A-label back.  

That may be just a different perspective on what Patrik is
saying, but reserving PVALID (or some replacement word) for
characters that can be represented _exactly_ in labels and
having a different kind of category for those that map feels as
if it would be helpful in establishing that distinction.  

My guess is that it isn't "PVALID" versus something else,
however, since, if we do this, both would be valid under the
protocol... so we may have to start thinking about terminology.

   john







More information about the Idna-update mailing list