[llvm-dev] RFC: Introducing an LLVM Community Code of Conduct

Joachim Durchholz via llvm-dev llvm-dev at lists.llvm.org
Tue Oct 13 04:37:43 PDT 2015

Am 13.10.2015 um 12:52 schrieb Renato Golin via llvm-dev:
> On 13 October 2015 at 02:42, Chandler Carruth via llvm-dev
> <llvm-dev at lists.llvm.org> wrote:
>> On behalf of the board of the LLVM Foundation, I’d like to start the process
>> of introducing a formal code of conduct for the community.
> Hi Chandler,
> I agree we need such guidelines, and I truly believe you're not trying
> to promote which hunts, so maybe some of the text needs to be toned
> down a bit.
> Following some of the comments, it would be good to see the
> annonymised result of your research, if at all possible. I understand
> some descriptions might uniquely identify people. :)
> I also agree that some descriptions are too verbose, and trying to be
> exhaustive, whereas any list will be incomplete, we might just as well
> be short and meaningful. I don't think we need to be legal here, since
> LLVM spans across so many countries, you won't be able to be free from
> all of their idiosyncrasies.
>> - *Be careful in the words that you choose.* We are a community of
>> professionals, and we conduct ourselves professionally. Be kind to others.
>> Do not insult or put down other participants. Harassment and other
>> exclusionary behavior aren't acceptable.
> This sums up well. The rest is just outlining behaviours that occur
> with extremely low frequencies on this community, some of which I have
> never seen.
> Another important thing to mention is the use of the English language
> as an international communication protocol.
> I haven't done the maths, but I believe most of the LLVM developers
> don't speak English as their native tongue. Moreover, even between
> English speaking countries, what's normal for some, is offensive for
> others. This is at the core of any open source community and cannot be
> ignored.
> Therefore, we need a special mention for people that do get offended
> to first, check with the offender to make sure there is intentional
> harm, then chose between trying to solve on your own *privately*, or
> reporting the case, as outlined by this document. Even the most
> "obvious" offences may end up as just using the wrong English word.

I have seen several occasions of wrong English word, but this was never 
a problem in itself.
I guess non-native English speakers already know how easy it is to 
misunderstand words, so they account for that already.
English being as diverse culturally as it is, I guess even native 
English speakers already know this.

I wouldn't put that into the code of conduct but in the Tips & Tricks 
for Successful Community Interaction text.

> Cultural differences are important, and if we start kicking people out
> you may end up with contributors from a single "social style", say,
> only "politically correct" people. Of course, this is a balance, and
> the threshold is arbitrary. That's why I propose to encourage people
> to resolve matters personally first, and try to identify malice before
> reporting.

Just curious: Do people ever do it the other way around?

>> If you believe anyone is in physical danger, please notify appropriate law
>> enforcement first. If you are unsure what law enforcement agency is
>> appropriate, please include this in your report and we will attempt to
>> notify them.
> Isn't this a bit too much? I mean, was there *ever* a case that people
> were in any physical danger?

I'm aware of such things happening in social projects, but not in any 
technical one.
Maybe because perceived personal stakes are higher in a social context.

> I think mentioning it once is ok, but you already have a bullet below
> on this topic, I don't think we need a whole paragraph for that, too.

In particular, dwelling on the topic much gives an impression of LLVM 
being a community that needs a detailed response to physical threats.
I wouldn't want to give that impression even if that were the case, to 
avoid attracting the types that like a more violent community.

>> Once the working group has a complete account of the events they will make a
>> decision as to how to respond.
> How is this decision made? Will the accused be involved in all cases?
> Can the accused defend him/herself before being judged?
> I'd normally assume so, but the following paragraph worries me:
>> Only permanent resolutions (such as bans) may be appealed. To appeal a
>> decision of the working group, contact the LLVM Foundation Board at
>> board at llvm.org with your appeal and the board will review the case.
> Here, you're separating what you can do with and without the right of
> defence, as an executor.
> It may be a way to avoid bureaucracy and get to the point, but it also
> gives a lot of power to a foundation whose members were not voted, nor
> chosen by the community in the first place.
> I don't believe this is in the best interest of the community, since
> the community has no say in it, nor control over it.
> If at least the reporting committee's members were interested
> volunteers, or voted representatives, then it'd be a matter of
> replacing them if they abuse of power. Currently, we have no such
> mechanism, nor guarantees that due diligence will be equivalent for
> members of the public versus members of the Foundation.
> I'm not saying it will, I'm just saying that if if does happen, we'll
> have no control over it. I'd like to avoid getting there in the first
> place.

Ultimately the buck will stop somewhere.
If it's delegated back to the community via voting, it will promote 
election campaigns - remember: Where there's an office to hold, there's 
somebody who wants to hold it just to be important, not because that 
somebody is particularly suited for the job.

As long as the committee remains unimportant, things will be fine.
As soon as the committee becomes important, the LLVM project has already 
degenerated into a snake pit, and is likely to fork anyway.

So it may be undemocratic, but people can easily fork the situation, 
something that you can't do in a state.

Just my 2 cents.


More information about the llvm-dev mailing list