[csw-maintainers] Friction induced by our package release model

[Maciej (Matchek) Blizinski]

I've seen that it quite often happens that a package or a set of
packages is being worked for quite a long time with no review from the
release manager.  If there is a number of maintainers involved, they
usually discuss, and achieve consensus about each issue.  When it's
all done, packages are submitted for release, and - like a lightning
out of a clear sky - the release manager says "no", and nobody
understands why.  It's a rare case that the raised objections are
clear problems with a package, such as a missing dependency.  They are
usually matters subject to opinion, aesthetic or otherwise.  Consensus
driven communities work these things out, and once consensus is
achieved, they don't come back to the same discussion right away.  If
there's a person who picks up a topic again and tries to undermine the
consensus without a good reason, the person is likely viewed as
poisonous.

As it happens, our workflow leads to exactly this.  Release manager
doesn't look at packages until they are submitted, and they are
submitted after a consensus between maintainers is reached.  If the
release manager objects, he's up against an already established
consensus.

It's not that the objections raised are unreasonable - they might be
well reasonable.  The problem is that they are raised at a wrong time.

Have other people noticed this?  Do you think it could at least partly
explain the excessive amount of friction in our community?