On 12/05/2021 02:17, John C Klensin wrote:
(trying again. Again, my apologies -- explanation on request)
--On Tuesday, May 11, 2021 22:51 +0100 Stewart Bryant
<stewart.bryant@xxxxxxxxx> wrote:
On 11 May 2021, at 20:24, John C Klensin <john-ietf@xxxxxxx>
wrote:
...
The AD can edit the text and provide a solution that works.
And which has IETF consensus?
In the majority of cases the resolution of an erratum is a
simple obvious fix.
Indeed. And in even a larger majority of cases, it is obvious
that a reported erratum either correctly identified the problem
or resulted from significant confusion on the part of the
reporter (which might be a separate problem).
I'm concerned about the other cases, but your comment suggests
a way to streamline processing even more than the current IESG
Statement suggests. Consider the following steps after
submission:
(1) If the reported error is in a document produced by a WG
that is still active, the report goes to them. If they decide
they want to deal with it, they can either verify or reject or
just identify it as "Held for Document Update" (and,
potentially, decide to go to work on it). If they cannot reach
consensus (either on reviewing or on what action to take, the
status should clearly be "Held for Document Update". Other than
some quick administrative actions, neither the RFC Editor nor
any IESG member need to be involved. In any event, modulo the
usual appeal provisions, their decision is final.
(2) The RFC Editor, in consultation with the author(s), make a
determination of whether the matter is a simple editorial issue
and, if so, whether the proposed fix is correct. If the answer
to both questions is "yes", then the report is verified.
Otherwise, go to the next step.
(3) Some relevant AD (obviously the one who worked on the
document earlier if there is such a person) consults with
authors, present or former WG chairs, and, at their discretion,
other ADs and anyone else who seems relevant. They make a
determination as to whether the problem is correctly identified
and the proposed fix is correct, simple, and obvious. If those
who are consulted can reach consensus, the report is either
verified, rejected, or classified as "held for...". If they
cannot agree, the matter is sufficiently controversial that
"Held for Document Update" is the correct status
classification. It would be a good idea to announce that
decision to the community in case someone sees issues the
review team missed or otherwise objects but I'd see that as a
monthly or per-IETF-meeting summary, not noise on the IETF list
(or even a relevant ex-WG list) per reported error.
John,
Mostly, I agree but would suggest two wrinkles.
One is that when the WG who produced the RFC is not longer active, there
may be a successor WG which is competent to reach consensus on the
erratum, e.g. ipsecme, lamps, and I think that whether or not such a
group exists is a judgement call for an AD from the area, so I would
insert such a step - pass to Area AD to see if another WG can step in -
before going to the RFC Editor (whom I would not see as necessariy
having skills in that field).
Second, you say that 'Hold for Document Update' makes it painfully clear
that there is no IETF consensus on the problem. I disagree. The Editor
page says that this status says that it is not a necessaary update, may
be considered for a future update which to me has always implied that
this may not be an error at all, rather out of scope at present, a
request for a functional change, such as to make a product compliant to
the RFC:-). SO there may be a clear consensus, not right at this time.
Tom Petch
So, if things are simple and obvious, they get dealt with very
quickly and efficiently. Or, if they are not, they get
classified as "Held for Document Update", which makes it
painfully clear to the reader that there is no IETF consensus
on the problem and solution. It also puts the decision as to
what to do about the report when there is no conensus back into
our normal process rather than using errata reports for
non-obvious problems or solutions as a means of changing
documents with little or no community input.
There is obvious potential for DoS attacks in both the above
and in the IETF Statement as it now stands, but our current
procedures for dealing with people who are disruptive ought to
be perfectly adequate to handle them.
best,
john
.
