Re: [EAI] I-D Action: draft-ietf-eai-mailinglistbis-00.txt

[Frank Ellermann <[e-mail]>]

On 23 November 2011 17:32, John Levine <[e-mail]> wrote:

>>|  For more information, see [listheaders].
>>Pointer apparently AWOL.
> Bug in xml2rfc.  They know about it.

Yes, I saw your messages on the xml2rfc list, but as user of
the online xml2rfc tools I can't tell how to fix your offline
installation.  For the online tool if I don't like the bibxml
output I just copy the bibxml snippet and then tweak it until
I'm satisfied. ;-)

>> Please add an example.  At the moment VERP is "specified" in
>> Wikipedia, maybe this draft could explain it for the IETF.

> It's per-recipient bounce addresses that embed the recipient's
> address.  For this doc, I don't see any need to describe it in
> more detail.

Mostly I have forgotten all VERP and BATV details, but there are
"local part length" dragons, IIRC.  For SPF "local part" simply
*is* the dragon, otherwise SPF would be EAI-ready, and all weird
security considerations for alternate addresses will be obsoleted
by EAIbis.

>> Please strike or replace this IRIbis detail, for the purpose
>> of "legacy" (= existing) list tools only the "cannot punycode"
>> interpretation makes sense, other <iregname> constructs, i.e.,
>> stuff outside of LDH DNS, are not relevant for mailing lists.

> But this isn't about legacy list tools.  It's about how list
> tools might change to handle EAI mail.

Where does using a percent-encoded instead of a punycoded <ihost>
help wrt mailing lists?  So far I think you can either keep UTF-8
"as is" or "get it right" (= 3987 + 5890).

-Frank