> -----Original Message----- > From: Stanislav Malyshev [mailto:smalys...@gmail.com] > Sent: Sunday, January 24, 2016 7:51 AM > To: Andrea Faulds <a...@ajf.me>; internals@lists.php.net > Subject: Re: [PHP-DEV] Re: [RFC] [Re-proposed] Adopt Code of Conduct > > Hi! > > > This is a question I'm wondering about as well. It all seems pretty > > good, but I wonder if, for example, the lists of unwelcome behaviour > > and discrimination characteristics are sufficiently complete. > > They are not supposed to be complete.
Actually, the way I read the golang CoC this isn't just a list of examples like the Contributor Covenant - but rather, the full list of issues that are considered to be forbidden. The penalties are supposed to apply only to people who break those: " These actions are explicitly forbidden in Go spaces:" ... " If you conduct yourself in a way that is explicitly forbidden by the CoC, you will be warned and asked to stop. If you do not stop, you will be removed from our community spaces temporarily. Repeated, wilful breaches of the CoC will result in a permanent ban." As personally I think even the current list is too wide unless constrained by clear (and fairly limiting) definitions for what constitutes harassment, bullying and demeaning behaviors - I think it's critical that this list won't be extensible-at-will of whomever gets to be in charge of enforcing it. > Again, we're not writing a penal > code. If somebody does something clearly destructive but not in the laundry > list, we'd have to deal with it and we'll find the way to. That's true, but I think that somewhat goes against the point of having a 'penalizing CoC'. The list of violation is supposed to be detailed and exhaustive. As I said numerous times, I would much rather deal with the 99% and keep the 1% (which is probably even less than that) undefined. Defining these penalties complicates things substantially: 1. Instead of picking people who are good mediators, we need to pick people that also have good investigative / judgmental skills. I would say that at least in my experience, it's uncommon to have these qualities in the same person. 2. Potential penalties hovering above the otherwise positive CoC introduce negative vibes, which harm the positive vibes. IMHO, those arguing we must support the 1% are doing so at the expense of the 99%. There's no way to optimize for 100% of the cases. Given that we've had zero cases of the 1% (so far nobody presented any case that would be considered a violation of the 'Unwelcome behavior' in the golang CoC had it been in effect in PHP). > For "discrimination characteristics" it's even worse, as we'd be pretending to > have a complete list of human identity and background groups, which we can > not have, and each failure mode for it is worse than the other - either we > presume anything we omitted does not exist, or is not important enough, or > it's OK to discriminate against members of that group. As far as I can see this is more or less based on what's in the US law: http://www.eeoc.gov/laws/types/ Again, here too, I think it is in fact supposed to be exhaustive, with the exception of 'similar *personal* characteristic'. Would someone being a very bad coder, and consequently not getting source control access be considered discrimination? Would someone not willing to cooperate with a member or supporter of a US/EU designated terrorist group (say ISIS) be considered discrimination? It's a brave new world, these aren't theoretical situations. They can happen in real life. Zeev
