Hi Barry,
Clarifications in line...
On 19/10/2016 08:55, Barry Leiba wrote:
I broadly in favour of this change, but I have a few comments.
Thanks, Brian.
First, a minor suggestion on the text itself:
OLD:
The responsible AD should
still check for downrefs before sending out the last call notice, but
any need to repeat a last call if this has not been done is at the
discretion of the IESG.
NEW:
The responsible AD should
still check for downrefs before sending out the last call notice, but
if an undetected downref is noticed during last call or IESG review,
any need to repeat the last call is at the discretion of the IESG.
Yes, that reflects the intent and is more explicitly clear. I like the
change.
Second,
there are no related security
considerations.
That bothers me a tiny bit. A missed downref could have security
implications.
I agree, but I contend that this doesn't make it any more likely that
we'll miss a downref. In fact, this change is only operable when we
*don't* miss it -- it simply gives the IESG judgment on whether last
call needs to be repeated when we catch it. And the Security ADs will
certainly have a say in that, if they think that broader review of the
downref is warranted for security checking.
True. How about s/related/directly related/ ? But I certainly don't insist.
Third, I believe that in addition to this procedural change, there is a
little work needed on the ecosystem:
1. Make the downref registry an intrinsic part of the data tracker. I mean
that
each document listed at
https://trac.tools.ietf.org/group/iesg/trac/wiki/DownrefRegistry
would instead be tagged as 'downref allowed' in the tracker, with
appropriate GUI
support for the IESG to apply this tag.
2. Enhance idnits slightly to check this tag when it detects a downref.
A downref to a 'downref allowed' document would be a warning, and a downref
to a non-downref-allowed document would be an error.
I agree that those would be excellent changes, and I'll ask Ben, as
sponsoring AD, to send that request up to the tools team. I don't
gather than you're asking that the document be held for that, correct?
Correct; it's orthogonal.
The change you suggest above is in my working copy for the next draft
revision.
Thanks!
Brian
Barry