[llvm-dev] Resuming the discussion of establishing an LLVM code of conduct

Hal Finkel via llvm-dev llvm-dev at lists.llvm.org
Thu May 5 13:03:02 PDT 2016

----- Original Message -----
> From: "Renato Golin via llvm-dev" <llvm-dev at lists.llvm.org>
> To: "Chandler Carruth" <chandlerc at gmail.com>
> Cc: "llvm-dev" <llvm-dev at lists.llvm.org>
> Sent: Thursday, May 5, 2016 2:06:30 PM
> Subject: Re: [llvm-dev] Resuming the discussion of establishing an LLVM code of conduct
> On 5 May 2016 at 19:48, Chandler Carruth via llvm-dev
> <llvm-dev at lists.llvm.org> wrote:
> > In my opinion, the community is growing and growing up. We need to
> > take some
> > steps to ensure that our long standing tradition of professional
> > and polite
> > behavior continues.
> I'd like to emphasise the "growing up" point, and how it is not
> directly related to *having* a code of conduct.
> People grow up all over the world and they get to be very different
> in
> their beliefs, behaviours and style. Communities are no different.
> Say what you want about the Linux kernel community, but you can't
> call
> it immature. You can call the behaviour of some of its people
> immature, but the community itself is not by a long shot.

But there are reasonable people who will not interact with that community because they find that community's acceptance of offensive behavior unacceptable. I certainly don't want to see that happen here.
> Codes of conduct, as they come, are an expression of some people,
> mostly from the US, in response to an extreme prejudicial behaviour,
> and is related to the recent political climate in that country, more
> than anything else.
> Communities grow up every year in many places, they get to be decent
> and caring (like ours, like Fosdem) without a code of conduct. That
> very reason is a fact against the idea that we *need* a CoC.
> But I'm not arguing that we don't, either. I'm supportive of a code
> that is not overly US-centric and that doesn't curb real enthusiasm
> when it's accompanied by a different culture. The current wording is
> still too much on that side.

I don't understand what you mean by US-centric, perhaps you could elaborate?

My experience, working in the US at research-oriented organizations, is that a lot of our sensitivity around these issues has nothing to do with US politics, as such, but that we have a diverse workforce, with wide representation from different cultures and traditions. Layered on top of that are variances in English comprehension, social skills, personal history, and other factors. We try really hard to make sure that everyone feels welcome and comfortable, and that means maintaining a level of professionalism and respect that translates well across cultural boundaries. We work with various vendors, university research groups, etc. and many of these groups are just as, if not more, diverse (and in many different countries).

Right now, whether it is our own employees doing the work, or a vendor's employees, we can set an expectation that those employees interact with the LLVM community when doing LLVM-related development. Should the level of discourse in the LLVM community degrade from its current high standards, an employee might object that interaction with the community would make them uncomfortable. That's game-over for us.

Thanks again,

> Trying to protect your culture is fine, but encoding your culture as
> the "right culture" to be followed is not nice when there are people
> from all over the world here. By enforcing a strongly biased CoC
> (which the current version still is, US-centric), may give the idea
> for people to abuse of their new-found power.
> So, whatever list of people's choices and birth marks you want to put
> in the "protected list of people", there will always be others.
> Whatever "accepted behaviour" is described, there will always be
> others. Whatever "unaccepted behaviour", there will always be a grey
> area where that's not exactly what happened. Encoding the grey areas
> will make it much worse, not better.
> But it's when you mix that, with the banning power of an
> yet-undefined
> group of people without any mention on how they'll be formed or
> operate, that you're opening the code ripe for abuse from all angles.
> Most of the abuse, however, will come unintentionally, when the
> people
> in the committee will form a biased consensus and not allow a chance
> of defence. When people in the community will unintentionally gang-up
> on someone they don't like to force the code on them.
> If you think this wouldn't happen normally, then know that on our
> previous exchange I have been personally emailed by half a dozen
> people telling me how much the code would shut me down, even though
> publicly, people said they didn't think I was being harsh /
> disrespectful.
> I fear the new code will bring that. A simpler, less power enabling
> and less biased code would not give anyone that power, and would be
> far more inclusive than the current one.
> cheers,
> --renato
> _______________________________________________
> LLVM Developers mailing list
> llvm-dev at lists.llvm.org
> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev

Hal Finkel
Assistant Computational Scientist
Leadership Computing Facility
Argonne National Laboratory

More information about the llvm-dev mailing list