W3C home > Mailing lists > Public > uri@w3.org > November 2006

Re: I-D ACTION:draft-duerst-mailto-bis-03.txt

From: Frank Ellermann <nobody@xyzzy.claranet.de>
Date: Wed, 01 Nov 2006 14:53:08 +0100
To: uri@w3.org
Message-ID: <4548A6C4.429@xyzzy.claranet.de>
Cc: ima@ietf.org

Martin Duerst wrote (<ima.ietf.org>):

> There is a special proceduce with which you can request an exception
> from the 'no downreferences' rule. If I remember correctly, it was
> proposed by John :-).

The proposed experiment didn't make it.  And if RFC 3967 was published
before 4409... now I'm curious if I can reconstruct this:

pidtracker.cgi?command=view_id&dTag=11010 says that 3967 was approved
2004-07-23 and published 2005-01-04.
pidtracker.cgi?command=view_id&dTag=11926 says that 4409 was approved
2005-05-16 and published almost a year later.  And the Last Call was
2005-01-27.  The tracker didn't note the text of this Last Call, asking
GMaNe I get <http://article.gmane.org/gmane.ietf.announce/4673>

Nothing about 3967 and Down-Ref, the normative reference of 2821 is an
error.  The dependency tool finds RFC 2821, but it doesn't red-flag it
as Down-Ref.  Some IETF rules & rituals are stranger than "humming" :-)

> Can you say exactly what you want to happen with In-Reply-To?
> (ideally some proposed text?).

Your text is fine, you only forgot to list it together with the known-
to-be-harmless hname "subject=" in chapter 4.  Where you have a hname
"body=", which is known-to-be-not-harmless.  And you have no "cc=" in
chapter 4, it's IMO better than the dubious "to=".

Let's move the mailto details to another list - the URI list isn't my
favourite, finding 2822 experts like Bruce is simpler on the rfc822
list, but you mentioned that you don't read that in the "archived-at"
Last Call.  I can't set a Reply-To or Followups-To via GMaNe, so this
is only an X-Post, please remove EAI.

>> or even UTF-8, encoded with 2047 (or a CTE + Content-Type for the
>> insane body= construct).

> What would be the issue with this? This would be the right thing
> to do pre-EAI, and is shown in various examples in the draft.

ACK, I meant UTF-8 C3 80 misinterpreted as say windows-1252 C3 + 80,
converted to C3 83 + E2 82 AC, and at that point adding 2047 doen't
help anymore, =?utf-8?q?=c3=83=e2=82=ac?= or similar is unrelated to
the u+00C0 I want.

 [after the EAI experiment]
> at that time we would also add the syntax for fallback addresses,
> if we think it makes sense, yes?

Yes, apparently they intend to take square brackets or curly braces.

John said the former might be better than the latter, but wrt URIs
it's apparently equivalent, 1738 had both in <national>, and 3986
requires to percent-encode both - IPv? delimiters in an <authority>
don't affect mailto:, it has no <authority> and no <host>, it uses
percent-encoded <addr-spec>.

>> Then we'll also have IDNAbis, and folks are used to work with
>> normal IRIs, not this worst case "mailto:"

> Can you say exactly what you are alluding to?

When IRIs are used everywhere folks will also want to use them for
mailto:.  They won't start to use IRIs with mailto:.  My crystal
ball says (caveat - the same crystal ball claims that we all use
X.400 for more than a decade now, shortly after X.25 and X.75 have
replaced the last pockets of IP :-)

Frank
Received on Wednesday, 1 November 2006 14:00:40 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:25:10 UTC