Re: Formal Review Retrospective

From: Vinnie Falco (vinnie.falco_at_[hidden])
Date: 2024-11-10 22:07:14

On Sun, Nov 10, 2024 at 11:20 AM Peter Dimov via Boost <
boost_at_[hidden]> wrote:

> We're between a rock and a hard place here; on one hand, we suffer from
> a lack of volunteers who would want to manage reviews, on the other, when
> Klemens volunteers, people don't like the way he manages the reviews.
>

I think there is a little more to it than that:

* Review managers have varying levels of skill
* Long-timers have better than average review manager skills
* New review managers don't receive implicit knowledge
* There is no process for mentoring review managers

If we want review managers to get better we have to grow them not find
them. This means the old teach the young. And of course we must document
oral traditions so they can be learned on demand. Some of the long-term
work we are doing is to index and cross-reference ALL historical reviews
going back to 1999 so they can be studied and have machine learning tools
applied (which reviews were good, which were bad, can we learn anything
from each, and so forth). I can't even answer a simple task, to find an old
review that was very well written and informative.

When I saw the first review result write up for async-mqtt5 I was
disappointed, as it did not rise to the level of quality that I have seen
in years past (no need to belabor this point as Peter has already described
this in another post). And my response was to go to the review manager
privately, and offer assistance so an improved write up could be published.
Among other things I advised, to not name people directly (as this can come
across as hostile) and rather, to name the ideas instead. And other stuff.

We have at least two reviews scheduled, and I have concerns that our pool
of review managers is light on experience and knowledge of the review
process and its culture. If we want to avoid unsatisfying outcomes, then we
need to be proactive. It would be nice if a long-timer skilled in the
review process could volunteer to act as a mentor for the
review-manager-in-training. Peter Turcan can assist with transcribing
cultural aspects into the documentation. And we can all take a pause and
maybe consider how we can take active steps to improve outcomes instead of
criticizing them after the fact.

Thanks