MUST canonicalize according to RFC 5198?

[Stephane Bortzmeyer]

I know it is quite late now that the I-D is in AUTH48 but I'm bothered
by a recent (outside WG, it seems) addition.

RFC 5198 is a normative reference. draft-ietf-ippm-twamp-09.txt says:

  The protocol does not carry any information in a natural language,
  with the possible exception of the KeyID in TWAMP-Control, which is
  encoded in UTF-8 [RFC3629, RFC5198].

This was apparently added in response to IESG comments like Chris
Newman's remark in <https://datatracker.ietf.org/idtracker/ballot/2229/>.

Does it mean that a TWAMP implementation MUST canonicalize the KeyID
according to RFC 5198, section 2? I may be dense but it is not obvious
for me.
_______________________________________________
ippm mailing list
ippm@...
https://www.ietf.org/mailman/listinfo/ippm

[Henk Uijterwaal]

Stephane Bortzmeyer wrote:

> I know it is quite late now that the I-D is in AUTH48 but I'm bothered
> by a recent (outside WG, it seems) addition.
>
> RFC 5198 is a normative reference. draft-ietf-ippm-twamp-09.txt says:
>
>   The protocol does not carry any information in a natural language,
>   with the possible exception of the KeyID in TWAMP-Control, which is
>   encoded in UTF-8 [RFC3629, RFC5198].
>
> This was apparently added in response to IESG comments like Chris
> Newman's remark in <https://datatracker.ietf.org/idtracker/ballot/2229/>.
>
> Does it mean that a TWAMP implementation MUST canonicalize the KeyID
> according to RFC 5198, section 2? I may be dense but it is not obvious
> for me.

Al, Keynam, others.

Since we have another bug to be discussed by the WG, can you also
address this one?

Henk

--
------------------------------------------------------------------------------
Henk Uijterwaal                           Email: henk.uijterwaal(at)ripe.net
RIPE Network Coordination Centre          http://www.amsterdamned.org/~henk
P.O.Box 10096          Singel 258         Phone: [phone]
1001 EB Amsterdam      1016 AB Amsterdam  Fax: [phone]
The Netherlands        The Netherlands    Mobile: [phone]
------------------------------------------------------------------------------

Is one of the choices leaving the office open?
                                       Alan Greenspan on the next elections
_______________________________________________
ippm mailing list
ippm@...
https://www.ietf.org/mailman/listinfo/ippm

[Al Morton]

At 02:25 AM 9/18/2008, Stephane Bortzmeyer wrote:

>I know it is quite late now that the I-D is in AUTH48 but I'm bothered
>by a recent (outside WG, it seems) addition.
>
>RFC 5198 is a normative reference. draft-ietf-ippm-twamp-09.txt says:
>
>   The protocol does not carry any information in a natural language,
>   with the possible exception of the KeyID in TWAMP-Control, which is
>   encoded in UTF-8 [RFC3629, RFC5198].
>
>This was apparently added in response to IESG comments like Chris
>Newman's remark in <https://datatracker.ietf.org/idtracker/ballot/2229/>.
>
>Does it mean that a TWAMP implementation MUST canonicalize the KeyID
>according to RFC 5198, section 2? I may be dense but it is not obvious
>for me.

Stephane,

Chris asked to include the UTF-8 reference in this section,
and in a subsequent e-mail the 5198 reference, and both add
clarity to what "UTF-8" is and its use in protocols. See:
https://datatracker.ietf.org/idtracker/draft-ietf-ippm-twamp/comment/84276/?

Otherwise, the wording of this text in this section is identical
to OWAMP's:
>9.  Internationalization Considerations
>
>    The protocol does not carry any information in a natural language,
>    with the possible exception of the KeyID in OWAMP-Control, which is
>    encoded in UTF-8.

There are no RFC 2119 requirement terms in section 9 of either OWAMP
or TWAMP. KeyID is only mentioned as a "possible exception".

IMO, inserting references (normative or informative)
does not add requirements unless the text specifies it
(and the level of requirement, using RFC 2119 terms).

Al

_______________________________________________
ippm mailing list
ippm@...
https://www.ietf.org/mailman/listinfo/ippm