From: Paul Overell (paulo@turnpike.com)
Date: Wed Sep 06 2000 - 08:42:17 CDT
In article <200009061027.LAA16279@clw.cs.man.ac.uk>, Charles Lindsey
<chl@clw.cs.man.ac.uk> writes
>Here are the issues on which I need some guidance. Please reply by next
>Thursday, Sept 14th, at the latest.
>
>-------------------------------------------------------------------------------
>
>1. Signed headers
>
>A. Do we proceed on the basis of draft-lindsey-usefor-signed-00.txt?
>
> [YES / NO]
ABSTAIN
>
>B. In the event the answer to A is YES, do we canonicalise dates as days since
> Jan 1st 1970 (in which case a time hh:mm:60 canonicalises the same as
> hh:mm+1:00) or do we canonicalise dates using date-time syntax (in which
> case hh:mm:60 and hh:mm+1:00 are regarded as different). The proposed
> wording of the two cases is at the end of this message.
>
> [DAYS since 1/1/1970]
> [date-time syntax]
>
date-time syntax
Avoids introducing yet another date format. Avoids having to
refer to an off-line reference. Avoids the leap-second bug and
potential security issue.
>
>2. "ought"
>
> Do we proceed on the basis of using "ought" as the word to indicate best
> practice in "social" cases where interoperability would not be affected,
> but the utility of Usenet to its users would. If this is agreed, then I
> will propose wording to define the usage, and then go through the document
> indicating where "SHOULD" should be changed to "ought".
>
> NOTE: Last time we were divided on this issue, but since then the IETF
> high-ups have indicated that using SHOULD for these cases is wrong, and they
> won't let us define "OUGHT". So we have not much choice, really.
>
> [YES / NO]
NO
I fail to see the distinction between "utility of Usenet" and
"interoperability". Are we absolutely sure that SHOULD cannot
be used?
>
>
>3. Mail-Copies-To
>
> Do we proceed to define the Mail-Copies-To header?
> The alternative would be to do nothing about this, or to include some
> alternative mechanism. Mail-Copies-To would, of course, have to be defined
> pretty much in conformance with its present informal usage.
>
> NOTE: Last time we voted 7:7 on this issue, so I am asking the question
> again in the hope of getting a clear result.
>
> [YES / NO]
>
YES
>
>And here are the promised wordings for the date canonicalizations.
>
>[1st alternative text]
>
> 2. Any date-time occurring in a Date, Resent-Date or Expires header
> (but not in any other header) is converted into the number of
> seconds since the start of January 1st 1970 UTC, ignoring any leap
> seconds and expressed as a decimal number without leading zeroes.
>
> NOTE: POSIX-compliant systems can implement this simply by using
> the POSIX 'mktime' routine [POSIX] with the TZ environment
> variable set to 'GMT'. Observe that the effect is to treat "31
> Dec 2000 23:59:60 +0000" (which is a legitimate date-time as
> defined by [MESSFOR]) as being identical to "1 Jan 2001 00:00:00
> +0000".
Does such implementation detail really belong here?
>[Can someone give me a reference to the proper POSIX document?]
>
>[2nd alternative text]
>
> 2. Any date-time occurring in a Date, Resent-Date or Expires header
> (but not in any other header) is converted to UTC and written as a
> date-time in the format
> 07 dec 2000 23:59:60 +0000
> Note absence of day-of-week, leading zero included in day, month-
> name in lower case, seconds included, year as four digits, and
> timezone preceded by a "+" as opposed to a "-", and observe that
> any FWS will be removed in the next step, giving
> 07dec200023:59:60+0000
>
> NOTE: Observe that the effect is to treat "31 Dec 2000 23:59:60
> +0000" (which is a legitimate date-time as defined by [MESSFOR])
> as being different from "1 Jan 2001 00:00:00 +0000".
>
As it should, they are different times.
>
>
Regards
-- Paul Overell T U R N P I K E