Re: [Emailcore] WGLC: draft-ietf-emailcore-rfc5322bis-03

[Alexey Melnikov <[e-mail]>]

Hi all,
(With my chair hat on.)

On Thu, Aug 4, 2022, at 1:45 PM, John C Klensin wrote:
> --On Thursday, August 4, 2022 09:20 +0200 Julien ÉLIE
> <[e-mail]> wrote:
>
>> Hi all,
>> 
>>>   |>    Though "-0000" also indicates Universal Time, it is
>>>   used to |>    indicate that the time was generated on a
>>>   system that may be |>    in a local time zone other than
>>>   Universal Time and that the |>    date-time contains no
>>>   information about the local time zone. |
>>>   |I probably still wouldn't...  I don't recall *ever*
>>>   |seeing "-0000".
>>> 
>>> Me neither, consciously.  Nor do i believe that most
>>> (practically, all) people ever will put value on that "-".
>> 
>> FWIW, dates with "-0000" often appear in the (Netnews)
>> Injection-Date header field generated by news servers when
>> injecting an article.
>> As well as in the Date header field when not provided by the
>> poster (this is not a mandatory field).  The news server adds
>> one, not knowing the local time zone of the author of the
>> article, and thus using the "-0000" syntax.
>> 
>> I even have an example from a well-known contributor of this
>> WG :-)
>> 
>> Message-ID: <t9d3rt$29g$[e-mail]>
>> Date: Mon, 27 Jun 2022 20:27:41 -0000 (UTC)
>> Injection-Date: Mon, 27 Jun 2022 20:27:41 -0000 (UTC)
>> 
>> 
>> Or when the Date header field is provided:
>> 
>> Message-ID: <tc6fmu$2q025$[e-mail]>
>> Date: Sun, 31 Jul 2022 19:55:42 +0200
>> Injection-Date: Sun, 31 Jul 2022 17:55:42 -0000 (UTC)
>> 
>
> So, just to be sure I understand, in the first case "-0000" is
> being used for "unknown for the actual sender/ originator/
> injector" above, with "(UTC)" the time as supplied by the
> Server.  Correct?
>
> Comment: 
>
> I believe there are three separate questions here:
>
> (1) Is it desirable to change 5322bis to allow a time zone to
> express non-integral minutes?  From the discussion I think the
> answer is "no", if only because these are supposed to indicate
> actual time zones rather than being, e.g., precise indicators of
> local time.

Indeed. Extending the format is a non starter.

> (2) Should "UT" be taken off the "obs-*" list and added back
> into valid current syntax to allow systems to be clear about
> what is intended, rather than having to guess at the intended
> semantics of "+0000" and "-0000".  I still feel that might be a
> good idea, but I'm not seeing much traction for it.

Right.

> (3) Recommendations about what time zones should be chosen for a
> given message, either in the "Date:" header field or in the
> various other fields, notably trace fields, in which time-stamps
> are applied and/or what receiving systems should do about them.
> That seems to be an A/S subject, not a 5322bis/5321bis one.
> Whether we have anything useful to say is another matter.

Agreed.

> Co-chairs: at what point will you step in on this and either
> declare consensus or issue a consensus call, so that this
> discussion does not go on forever in terms of 5322bis?

My reading of the mailing list agrees with your assessment, so at this point this discussion is closed.
(If people disagree about the discussion being closed, they at least must change the subject, as this is no longer a RFC5322bis issue.)

> Do you
> want to open a ticket about a time zone discussion in the A/S,
> or do we just move on?

I will open a ticket for A/S, but as you said, it is not clear whether we will do any textual change in relationship to it.

Best Regards,
Alexey