llvm dev - Oct 2015 - RFC: Introducing an LLVM Community Code of Conduct

On Wed, Oct 14, 2015 at 6:32 AM Krzysztof Parzyszek via llvm-dev <
llvm-dev at lists.llvm.org> wrote:
> On 10/14/2015 8:22 AM, Joachim Durchholz via llvm-dev wrote:
> >
> > Stop possible future harrassment before it can happen.
> >
> > It's a legitimate concern actually.
>
> CoC is not going to stop or prevent any harassment.

Do you have any data to back up this claim? I don't have any *data* that
says anything conclusive about whether a code of conduct will stop or
prevent harassment in an open source community -- there are relatively few
such communities and not much time in order to collect such data.

However, several events and communities have reported that their problems
with these issues (and they had more problems largely due to being larger
communities) decreased.

And all of the expert recommendations on how to reduce or prevent these
things suggests having a code of conduct precisely *because* it defines
these things and spells everything out clearly.


> It only defines
> what behaviors are considered "harassment", and that is something
we all
> seem to have a good grasp of anyway.

There have been numerous research studies here, and peoples' views on what
constitutes this varies *very widely*. This is especially true across
cultures. I can try to dig up links to some of them if that is useful to
you.
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20151014/a4c63f35/attachment-0001.html>

On 10/14/2015 9:43 AM, Chandler Carruth wrote:
> On Wed, Oct 14, 2015 at 6:32 AM Krzysztof Parzyszek via llvm-dev
> <llvm-dev at lists.llvm.org <mailto:llvm-dev at
lists.llvm.org>> wrote:
>
>     CoC is not going to stop or prevent any harassment.
>
>
> Do you have any data to back up this claim? I don't have any *data*
that
> says anything conclusive about whether a code of conduct will stop or
> prevent harassment in an open source community -- there are relatively
> few such communities and not much time in order to collect such data.
No, I don't.  I base my opinion on the notion that people don't do bad 
things not because they are illegal, but because the things are---bad. 
The "do unto others" rule is widely recognized for its universal value
and for most people it comes naturally.

> There have been numerous research studies here, and peoples' views on
> what constitutes this varies *very widely*. This is especially true
> across cultures. I can try to dig up links to some of them if that is
> useful to you.
No doubt that in the large enough group of people, these views will 
vary, however the LLVM community is not simply a set of people from the 
larger population.  It concentrates people with certain common 
background (type of education, type of profession, etc.), and with that 
background comes a certain common mindset.  In the time that I have 
participated in the mailing list discussions, I have not seen examples 
of behaviors that would suggest imposing a CoC.

I agree that a CoC for in-person meetings (such as conferences) may be 
needed, but they should also be crafted specifically to address problems 
that are pertinent to the circumstances.  For mailing list discussions I 
see it as unnecessary for reasons I have explained in other emails.

-Krzysztof

-- 
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, 
hosted by The Linux Foundation

On Wed, Oct 14, 2015 at 7:59 AM Krzysztof Parzyszek <kparzysz at
codeaurora.org>
wrote:
> On 10/14/2015 9:43 AM, Chandler Carruth wrote:
> > There have been numerous research studies here, and peoples' views
on
> > what constitutes this varies *very widely*. This is especially true
> > across cultures. I can try to dig up links to some of them if that is
> > useful to you.
>
> No doubt that in the large enough group of people, these views will
> vary, however the LLVM community is not simply a set of people from the
> larger population.  It concentrates people with certain common
> background (type of education, type of profession, etc.), and with that
> background comes a certain common mindset.

I believe you are incorrect about the commonalities of our technical
knowledge having any influence on whether we share a common (often
cultural) understanding of what is or is not going to constitute
harassment.

However, this is getting us off topic. I won't continue to debate this
point on the list.

>   In the time that I have
> participated in the mailing list discussions, I have not seen examples
> of behaviors that would suggest imposing a CoC.
>
To repeat what has been said several times in this thread:

This is not expected to be a dramatic change of anything. The community has
already been holding itself to exactly the kind of standard we are putting
into the document for many years.

This is a *preventative* measure and not about specific examples, so
arguing their have not been examples doesn't make much sense.
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20151014/0ef40413/attachment.html>

On 10/14/2015 9:59 AM, Krzysztof Parzyszek via llvm-dev
wrote:
>
> No, I don't.  I base my opinion on the notion that people don't do
bad
> things not because they are illegal, but because the things are---bad.
> The "do unto others" rule is widely recognized for its universal
value
> and for most people it comes naturally.
Let me add one more thing to this.

I believe that people are more committed to certain principles when they 
see them as their own.  An honor code works very well in certain 
communities, where doing the honorable thing is of high value.  I belive 
that the LLVM community is of that kind.  Imposing an external rules of 
conduct deflates these values, at it is no longer up to the individuals 
to uphold it.

I see that there is support for the CoC in this thread, so I'm not 
really seeing a chance that we can get away without one.  At this point, 
I just wanted to explain my views and concerns about it.

-Krzysztof

-- 
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, 
hosted by The Linux Foundation

Am 14.10.2015 um 16:59 schrieb Krzysztof Parzyszek via
llvm-dev:
> No, I don't.  I base my opinion on the notion that people don't do
bad
> things not because they are illegal, but because the things are---bad.
> The "do unto others" rule is widely recognized for its universal
value
> and for most people it comes naturally.
This is true as long as a social group is coherent. It starts to fail as 
the group becomes larger:

1) Some people accept more rudeness and think they can be more rude than 
is acceptable to other community members. These people follow the "do 
unto others" rule, but they're making the place unfriendly towards
those
who cannot stand that level of rudeness and will leave. Let this happen 
a few times, and over the years, you'll see a downwards spiral.

2) The human mind is very, very clever about finding justifications. The 
offender will continue to find justifications for his behaviour, and 
honestly believe them, even if nobody else does - if the offender is 
made aware of the fact, he'll simply claim that "majority does not make
right" and *still* feel justified.

3) The "do unto others" rule is not universally accepted, exceptions 
exist. Most notably, they do not aply to people "not in my group".
Some
people simply define their group, and consider "everybody too dumb to 
read the docs" an outsider and harrass them mercilessly. When their 
behaviour is challenged, they fall back to (2), such as "these idiots 
shouldn't be in this group, they can return once they have started 
learning".
> No doubt that in the large enough group of people, these views will
> vary, however the LLVM community is not simply a set of people from the
> larger population.  It concentrates people with certain common
> background (type of education, type of profession, etc.), and with that
> background comes a certain common mindset.
Sorry, but: no, I wouldn't take that as a given.
I have seen geniuses with totally inacceptable social behaviour.
And not all of them have the discipline to restrain themselves from 
participating in a community.

You might not have met these because your peer groups have been lucky 
enough to be self-selected, but it does not work in general.