[Development] QUIP 12: Code of Conduct

[Edward Welbourne]

Jason H (24 October 2018 17:09)
> - Is "Sceintific racism" actual racism or just statistics?

If it's racism, it's racism, however qualified.
Extrapolation from populations to individuals misuses statistics.
It isn't scientific, it just abuses tools lifted out of science.

> I really want to know where we are with James Damore because I thought
> his paper was well-researched with a scientific basis?

I had to look that name up.
While no source is unbiased, I'll take [0] as a tolerable source.
They do, at least, have a fairly solid understanding of what science is
(and isn't).

* [0] https://rationalwiki.org/wiki/James_Damore

Apparently he fails to understand the difference between very minor
statistical differences between broad populations and the details of
individuals.

Specifically: though the proportion of women who are good at certain
tech jobs might be marginally smaller than the proportion of men who
are, a recruiter who has the economic power of Google and commits to
recruiting equally should be able to do so, without compromising its
recruitment standards, provided there aren't *other factors* at play
that prevent it from doing so.  The crucial detail here is that Google
employs a tiny proportion of the population from which it could draw
recruits.  So half of Google's relevant technical staff - which is how
many women Google would need to hire to meet the given goal - fits well
within the available pool of suitably-skilled women.

Google would (in this hypothetical world) have to work a little harder
and pay a little more (but it's only a little, since the statistical
effect is quite small in fact) to find the women than to find the men,
but it's not short of applicants and Google, in particular, has
expertise in the field of selecting the best few from a plethora of
candidates - at least when it comes to pointing one at web pages.  The
fact that Google doesn't manage to hire equally many good women as men
in various tech positions *is* evidence that there are other factors at
play, aside from the scientific evidence of very minor differences in
aptitude (mostly stemming from differences in interest).

It is, furthermore, patently clear that the world does have other
factors that contribute to the gender divide in various jobs.  When
boards are dominated by men, it is no great surprise that women aren't
as widely represented in upper management, from the ranks of which most
boards are drawn, to take just one example.
But this is something of a digression.

> Having been interested in software from a very young age, and later
> specifically Open Source, one thing that appealed to me was that it
> was a meritocracy.

Well, many software practitioners at least aim to make software projects
meritocratic.  However, their ability to do so may be compromised by
social dynamics (and economics) in various ways.

> The best code survives, your code contributions are limited only by
> your code being the best.

If those evaluating how good something is are, unwittingly, operating in
an environment that some folk find hostile, those folk get driven off
and the evaluators fail to see how good their contributions would have
been, if they'd only felt at ease.  The aim of a code of conduct is to
avoid that.

I endure rudery from others moderately calmly, partly because I come
from a highly-privileged background that gives me the confidence to not
worry that the rudery will actually cause problems I can't handle.  I
prefer, and usually manage, to work in environments in which I and those
around me don't need to endure such rudery - partly because, while I can
endure it, I don't like it; but also because I don't want others to be
driven away, whose contributions I might welcome.

There may be bad codes of conduct out there; please don't let that put
you off trying to think about what a good code of conduct would look
like.  In particular, note that there are some "entrenched interests"
that don't seem to like codes of conduct; and they've taken pains to
talk up the misadventures of groups struggling to make them work.  Other
groups, garnering far less publicity, have bumbled along quite happily
for years with codes of conduct that seem to work fine.

So please don't just write off the code of conduct as a bad plan; try to
help us make a good code of conduct and a good process around it.
In particular, please at least read it before criticising it,

	Eddy.