Re: [idn] ZWNJ

["Adam M. Costello" <amc@cs.berkeley.edu>]

Roozbeh Pournader <roozbeh@sharif.edu> wrote:

> BTW, there are also many other needs for being able to retreive the
> original non-nameprepped name.

It will be nice to be able to use non-nameprepped names, but what do
you mean by "retrieve"?  If you mean recover them from ACE, that's too
much to hope for; ACE can store the nameprepped name and the mixed-case
info, but anything more would be difficult and inefficient.  If you mean
retrieve from DNS, I don't think that's helpful, because names retrieved
from DNS are very rarely seen by typical users.

I think what you really want is simply to be able to use non-nameprepped
names in application protocols, so that users can see them.  That will
be possible as soon the protocol adds explicit support for IDNs, so that
it no longer relies on ACE as the sole means of representing IDNs.

Maybe you were imagining that an application, after receiving an ACE
name in an application message, would do a DNS lookup to retrieve a
non-nameprepped version of the name.  That would not only increase the
load on the DNS, but might also return the wrong version of the name,
from the sender's point of view.  For example, if mail user agents did
that, everyone would see my email address as "amc@CS.Berkeley.EDU",
which I think is ugly.  If we want to allow users to see non-nameprepped
names, we should aim for the spelling chosen by the sender, not the
spelling chosen by the registrant.  That means the non-nameprepped
spelling has to be taken from the application message, not from the DNS.

AMC