At 06:59 PM 9/3/2004, william(at)elan.net wrote:
On Fri, 3 Sep 2004, Andrew Newton wrote:
> The debate over the MUST/SHOULD/MAY language with regard to compliance
> and publication of TXT and SPF2 record types does not seem to be
> getting the attention of the full working group.
>
> Despite lack of agreement on the outcome, there is agreement on the
> stated positions. In the words of Olafur:
> 1) Specify everyone to be compliant at all times.
> 2) Specify clearly full complaint state and tolerate non-compliant
> state during phase-in.
>
> Position 1 is predicated on the notion that a future RFC will deprecate
> the (use of the) TXT record in favor of the SPF2 record. Position 2
> demands that the TXT record be discouraged from use in the current
> -protocol document.
>
> Rather than wordsmithing the actual MUST/SHOULD/MAY language directly
> (and endlessly), we ask the participants state their preference on
> these positions.
I take middle position:
1.5) Specify everyone to be compliant at all times and tolerate
non-compliance as a matter of fact (current reality) for a given
period of time (next 4 years say).
Please be sure ALL such timeout dates are listed in drafts as a period of
time from when the draft is published as an RFC. Dates should NEVER be set
in a draft as pointing to a specific date on the calendar, nor from the
date a draft appears in the drafts repository.
We've run into issues before where "deadlines" were set and then acted upon
before drafts were approved. Please let's not ever do that again.