On 4/30/20 6:10 PM, Benjamin Kaduk wrote:
It also bothers me that the conversation around a particular last call
(other that private feedback) isn't all in some place that people can
easily read it and contribute to it.
I'm not sure in what way this is the case; it sounds to me like you are
saying that the last-call disucssion of a particular document is fragmented
across many places in the current experiment's setup.
Yes that's what I'm saying, though not entirely because of the current
experiment. This is a problem that has existed for decades, it's just
that the current experiment makes it a bit worse.
I honestly am not
sure what those places would be (that is to say, I have been assuming that
last-call@ captures all of it), since the directorate reviews sent via the
datatracker go to all three of the WG list, last-call@, and the directorate
list, and the last-call announcements set reply-to to last-call@. While I
have occasionally seen directorate review threads that drop one or two of
the lists, it is in my experience the exception, so I'd like to learn more
about how your perception differs from mine.
I'm not specifically concerned about the directorate threads, I'm
concerned about all Last Call comments. I think responses from
directorate members should get exactly the same exposure, be forwarded
to the same lists, etc. as Last Call comments from anyone else.
There also seems to be a disturbing trend toward over-fragmentation of
the IETF community conversation, for instance the now-common admonition
to move nearly every substantive discussion of potentially-broad
community interest to a separate narrowly-focused list to which one has
to explicitly subscribe. I do not think this serves IETF's interests,
and I think it too often serves to hide issues which should have wider
exposure. Separating last call announcements is probably among the
I think I understand your point here and sympathize that forcing a
discussion to split off from its current forum can result in the subsequent
discussion occurring among a too-narrow group of participants that does not
reach a conclusion reflective of the broader community's interests. But it
doesn't always do so, and there are many factors to be balanced in
analyzing what's best; I'm not sure that this current thread about the
particular ongoing experiment with last-call@ is the right place to get
into the broader discussion.
I agree with that; but it is one of my observations about the current
experiment and one of my reservations about making it permanent in its
current form.
least of the factors contributing to that over-fragmentation, but I
think it nonetheless contributes to the problem.
Here you say "separating last call announcements", which confuses me. In
what way are the *announcements* separated; my understanding is that they
continue to go to ietf-announce@, unchanged by the current experiment.
Yes, I was slightly confused there. Announcements go to ietf-announce
(IMO they should to go lastcall also), and the reviews/responses go all
over the place.
For the most part it doesn't bother me to not have last call
announcements cluttering up the conversation on the ietf list... except
that sometimes those conversations do need community-wide attention.
Again, the announcements should be going to ietf-announce@, so I'm not sure
how the ietf list comes in.
I should have said "last call conversations" or something similar.
Looking through my lastcall mail folder, what I see is that it's hard to
find the Last Call announcements for all of the area directorate reviews
that have been posted (with different subjects, and without links to the
drafts being reviewed). IMO the list would be MUCH more valuable if it
I see you saying that directorate review mails are being sent with
inconsistent subject lines and without links to the drafts.
Not quite; I'm concerned that it's difficult to get a sense of the
spectrum of support and/or concerns expressed in a Last Call. The
announcement goes to ietf-announce, some of the reviews go to lastcall,
some of the reviews go elsewhere. Why wouldn't we want them to all
show up and be reviewable in one place?
A related concern is that there are so many messages sent to lastcall
that there's no good way to even find the responses that are sent there.
My understanding is that the directorate secretaries are pushing reviewers
to have the datatracker send the review mail, which standardizes on the
message format and subject line (though apparently does not specifically
include a link to the document, which should be easy to rectify; please
feel free to file a ticket). So I'd be interested in knowing what
timescale your search spanned, and especially (though I acknowledge this is
not a trivial request) how the distribution has changed over time.
I haven't tried to continuously review this over time, just whenever I
wanted to follow a particular Last Call.
If I had to assign a number to the current implementation, I'd probably
give it a 1.5, but slight changes could make it a 4 or better.
I don't have a great sense for which of the potential changes discussed in
this thread would be needed to make it a 4 in your view. (I'm not going to
hazard a guess, either, since I'd expect to be wrong.)
I suspect I need to write up a strawman proposal that nails down some
details, because slight differences in implementation would make large
differences in participant experience.
Keith
