Nameprep and NFKC

Andrew Sullivan ajs at
Mon Oct 18 14:59:59 CEST 2010

On Sun, Oct 17, 2010 at 09:53:52AM -0400, John C Klensin wrote:
> In this particular case, I can't imagine any reasonable behavior
> other than to reject the string, presumably with an error
> condition, 

But that's this _particular_ case.  In an actual application, the
reasonable thing to do would be to kick the string back up to the
application's mapping layer, which either can put the string into NFC
or not.  But the error handling probably lives up there.  (This is
another reason why the RFC can read as it does: the _protocol_
requires certain things, but it doesn't constrain how things outside
the protocol layer behave.)


Andrew Sullivan
ajs at
Shinkuro, Inc.

More information about the Idna-update mailing list