[Ietf-not43] [Fwd: CRISP: Language/script-specific data sets?

[Ted Hardie]

> 
> FWIW, in the provreg WG protocol we're providing the ability to provision
> contact data in two character sets.  It would probably be a good idea if
> there were provisions to display one or both forms using the CRISP protocol.
> 
> -Scott-


Scott,
	I just looked through draft-ietf-provreg-epp-contact-05.txt,
and it looks like you have two different mechanisms which are related
to this problem.  One is the "lang" based negotation for the response
codes which is described in described in section 5 of
draft-ietf-provreg-epp-07.txt.  If I read that mechanism right, you
can at session start negotiate the "lang" attribute for the human
readable text associated with a response code.  This doesn't affect
the charset which SHOULD be UTF-8 unless a charset has been named.
(By the way, would this also affect things like the human readable
text in a statusValue?).
	The other mechanism would be charset related, and is described
in 2.3 (and again 2.4) of the contact draft:

  Individual and organizational names associated with a contact are
  represented using character strings.  These strings have a specified
  minimum length and a specified maximum length.  Individual and
  organizational names MAY be provided in both UTF-8 [RFC2279] and a
  subset of UTF-8 that can be represented in 7-bit ASCII depending on
  local needs.

	I read this to mean that they may be in *either*.  Does it in
fact mean that they will always be in in UTF-8 and may also be in the
subset of UTF-8 that can be represented in 7-bit ASCII?  If so, I
didn't see how which to present was negotiated in epp when both are
present.  Do you always send both when both are present?
			regards,
				Ted