Hello everyone,

Philippe Gagnon and I noticed that lazy consensus is occasionally being
applied too quickly to substantial Airflow changes, often before a broader
consensus has been built on the dev list. Recent examples [1][2] suggest
that while ideas are often hashed out on Slack or off-list, others may be
surprised when lazy consensus is declared without having had the
opportunity to participate in the discussion. No malicious intent - I
believe it results from people discussing thing in Slack or other places
long enough and deep enough to think that things are "obvious".

I’ve made similar mistakes myself in the (long)  past [3]. It is easy to
assume an outcome is obvious after lengthy private discussions, but the dev
list remains the essential place for transparent consensus building.

I would love to hear your thoughts on this and whether we all understand it
similarly and if we should adapt our current approach? Deciding things
through open discussion is a core part of why we are in the ASF.

BTW. If you are curious, my "consensus" from 7 years ago turned out to be a
bad idea. But over seven years, our current approach, including
dependencies, has become a much better and far more complex way to solve
the stated - and it constantly evolved (and will continue doing so).

Thanks,
Jarek Potiuk

[1] https://lists.apache.org/thread/fxxg0tgxd774vd9sl3w61mx69mmlk1gt
[2] https://lists.apache.org/thread/9yp0kwpld6ml2pzb8czpsk8zorgrqgh5
[3] https://lists.apache.org/thread/olyvhsvyw2dgqt0t64xmlk0og6to9fw5

Reply via email to