Re: rfc2822bis -- adding clarifying Sender: text

On 1/19/08 at 11:22 AM -0800, SM wrote:

>At 10:23 19-01-2008, Pete Resnick wrote:
>>Note: The transmitter information is always present. The absence of 
>>the "Sender:" field is sometimes mistakenly taken to mean that the 
>>agent responsible for transmission of the message has not been 
>>specified. The absence of the "Sender:" field merely means that the 
>>transmitter is identical to the author and is therefore not 
>>redundantly placed into a separate field.
>
>I prefer the above text to what is in Section 3.6.2 of the draft.
>
>The current draft (draft-resnick-2822upd-04) has:
>
>   "If the originator of the message can be indicated
>    by a single mailbox and the author and transmitter are identical, the
>    "Sender:" field SHOULD NOT be used.  Otherwise, both fields SHOULD
>    appear."

That text will still appear in the next draft. The above "Note" is in 
addition to that text.

>I viewed RFC 2822 as saying that we "must" put in a Sender if it's 
>different from the author.

Well, sort of. It's saying, "having a Sender field means that the 
author is different from the transmitter".

>If the author and the sender are the same, we "must not" put in the 
>Sender header as it would be redundant.
>
>On reading of the current draft, one may interpret that "SHOULD NOT" 
>as saying that it is recommended not to put in a Sender field unless 
>the circumstances dictate it.  To avoid "the circumstances" from 
>being stretched, we can either use a "MUST NOT" or else use your 
>note.

No. That's not how RFC 2119 MUSTs and SHOULDs are used. MUST NOT 
means that doing so is known to cause harm or interfere with 
interoperation. That simply isn't the case here. We can imagine cases 
where having redundant Sender information might cause confusion, but 
we say SHOULD NOT because there may exist valid reasons where the 
behavior is acceptable.

>I favor your note as it's a clarification of the issue that was 
>sought.  Here's a slightly modified version:
>
>Note: The transmitter information is always present. The absence of 
>the "Sender:" field is sometimes mistakenly taken to mean that the 
>agent responsible for transmission of the message has not been 
>specified.  This absence merely means that the transmitter is 
>identical to the author and is therefore not redundantly placed into 
>the "Sender:" field.

I'm fine with that as a bit clearer. But that's in addition to the 
normative information still in 3.6.2.

pr
-- 
Pete Resnick <http://www.qualcomm.com/~presnick/>
Qualcomm Incorporated - Direct phone: (858)651-4478, Fax: (858)651-1102