> If we ask new contributors to learn from the commit history, they are likely
> to
> find the inconsistency and feel confused.
simply decreasing complexity might be a reasonable choice here.
an honest question follows that i cannot confidently judge, but i do have an
intuition about:
is the net balance of the following two a clear positive?
1) total time spent on learning, communicating, arguing, and writing the
ChangeLog format messages for *every* commit (including the email excanges when
patches are rejected, let alone debating which "right" format it is)
2) the *decrease* in total time spent searching around in the repo history
and also considering that now we have `git log --patch --grep` and various
other tools.
PS: some of the second order effects could also be relevant here (e.g.
filtering contributors for personality type), but i doubt that would be a
conscious choice, and accordingly i don't see that discussed either.
--
• attila lendvai
• PGP: 963F 5D5F 45C7 DFCD 0A39
--
“When we become aware of how the school system is a conditioning agent to
instill in children obedience to authority, passivity, and tolerance to tedium
for the sake of external rewards, we begin to question school performance as a
metric of well-being. Maybe a healthy child is one who resists schooling and
standardization, not one who excels at it.”
— Charles Eisenstein, 'The More Beautiful World Our Hearts Know Is
Possible' (2013)
