Date: Thu, 23 Jul 2009 10:44:09 +0100
To: IMA <ima@ietf.org>
From: "Charles Lindsey" <chl@clerew.man.ac.uk>
Content-Type: text/plain; format=flowed; delsp=yes; charset=iso-8859-1
MIME-Version: 1.0
References: <mailman.13830.1247508102.4936.ima@ietf.org>
 <CAD7705D4A93814F97D3EF00790AF0B315FA6650@tk5ex14mbxc105.redmond.corp.microsoft.com>
 <4A5BABF8.4080900@isode.com>
 <CAD7705D4A93814F97D3EF00790AF0B315FA6AAF@tk5ex14mbxc105.redmond.corp.microsoft.com>
 <4A60AA0B.4000106@alvestrand.no>
 <CAD7705D4A93814F97D3EF00790AF0B315FCA179@TK5EX14MBXC104.redmond.corp.microsoft.com>
 <EA9664FBEBEB7127550C3D30@[192.168.1.110]>
 <CAD7705D4A93814F97D3EF00790AF0B315FCB1CA@TK5EX14MBXC104.redmond.corp.microsoft.com>
 <F2BC6EC973C4D97B22FB5FE1@p3.int.jck.com>
 <op.uxgqkfop6hl8nm@clerew.man.ac.uk>
 <8460C06800C257EDBC9EB95A@p3.int.jck.com>
Content-Transfer-Encoding: 8bit
Message-ID: <op.uxilbvrw6hl8nm@clerew.man.ac.uk>
In-Reply-To: <8460C06800C257EDBC9EB95A@p3.int.jck.com>
User-Agent: Opera Mail/9.25 (SunOS)
Subject: Re: [EAI] Rechartering
Precedence: list

Seems I sent my mail to John directly, rather than to the list. Here it  
is, together with his reply and my further views.

On Wed, 22 Jul 2009 16:58:11 +0100, John C Klensin <klensin@jck.com> wrote:

> --On Wednesday, 22 July, 2009 10:42 +0100 Charles Lindsey
> <chl@clerew.man.ac.uk> wrote:
>
>> What I had in mind was that the headers document (and probably
>> the smtp one as well) would fix the syntax of addresses with
>> alternatives (extra <...> and all). So any future work on
>> downgrade would have to live with that syntax. If that
>> syntactic change is to happen at all, it needs to happen NOW
>> because, as John points out, an awful lot of systems all over
>> the place are going to have to learn to allow it (at least to
>> the extent that they accept it and store it and/or pass it on
>> to further systems as an opaque bundle, even if they cannot
>> make use of it themselves).
>
> That is plausible.  However, the syntax change is a huge price
> to pay in terms of difficulty of changing systems and
> reeducating users unless it is going to support a downgrade
> mechanism that works and is significantly useful.   So I can't
> agree that changing the syntax to support downgrade without
> being sure that we have a downgrade model that we are also
> willing to standardize is a good idea.   One or the other --
> syntax change with a fully-fleshed-out and tested downgrade or
> no syntax change, no downgrade.

Sure. If we standardize the format, together with alt-addresses, then  
there must be a firm commitment to follow it up with a downgrade standard.  
But since the downgrade protocol is likely to need further tweaks, it  
could be better to leave it as Experimental until it has stablized.

I envisage that major pieces of software, such as servers, will be written  
with the downgrade part as an optional plug-in. The default plugin would  
be to "bounce everything" (which is a legitimate tactic under the present  
proposal). One up from that is a plugin that will recognize and forward it  
to the alt-address, but not much more.

Nobody wants to rewrite a full server each month, which is why the basic  
functionality must be fixed from the start. But upgrading your downgrade  
plugin (well, probably not monthly) is a task that could be more easily  
envisaged, and would be driven by the needs of the particular community  
you were trying to serve.

-- 
Charles H. Lindsey ---------At Home, doing my own thing------------------------
Tel: +44 161 436 6131                       
   Web: http://www.cs.man.ac.uk/~chl
Email: chl@clerew.man.ac.uk      Snail: 5 Clerewood Ave, CHEADLE, SK8 3JU, U.K.
PGP: 2C15F1A9      Fingerprint: 73 6D C2 51 93 A0 01 E7 65 E8 64 7E 14 A4 AB A5