Hacker News Reader

Comment by eznzt | Hacker News Reader

If you need a CoC to get rid of someone who harasses others what you have is a deeper problem

How else would you codify what is and what isn't acceptable behaviour?

Is it just arbitrary? Do you just depend on more people enforcing their views out to stop harassment? If so, then why not codify it?

> Is it just arbitrary?

Yes. Just as "kind" is. To the reporters and the officials, "saying somebody is wrong" is unkind, so it's clearly arbitrary.

The worst thing is when nobody is actually offended or hurt, and someone says your behavior could be considered offensive or insulting or whatever and takes action based on that. That's just someone wanting to exert their power for the wrong reasons.

That reminds me of that event when a person with screen name "Christian", of all things, filed an issue at VS Code around the, ah, New Year festivities, demanding to remove the Santa hat from the status bar because it's offensive. And the VS Code team promptly reacted, by agreeing, changing it for a snowflake (I still have no idea if it was ironic, or not, and whether they've noticed it later) and pushing out an emergency hot fix.

That did piss off quite a number of people, who filed much more hard-pressing (in their opinion) issues (like, some useful editor functionality being broken) and received precisely zero feedback.

You don't get to be the arbiter of who is "actually offended." I agree that there are people who take any amount of power and want to expert it for the wrong reasons, but that isn't an indictment against CoCs, it's an indictment of bad behavior.

That bad behavior is enabled by CoCs though, isn't it? If you leave room for abuse, you'll get abuse.

>> You don't get to be the arbiter of who is "actually offended.

That is exactly my point. Some committee member doesn't get to say "that may be offensive to someone" either. They need to wait until someone actually reports being offended.

As this whole debacles shows codifying doesn't remove arbitrary standards. Human social interactions are far too complex to codify in a concise way without massive edge cases. So all too often it just gives people a tool with which to better force their own arbitrary standards onto others.

As this blog post [1] lays out, this seems like a case where the CoC was poorly authored exactly for this reason. But this isn't a reason to not have a CoC, it just means this particular one was not done well.

[1]: https://www.fast.ai/2020/10/28/code-of-conduct/

There is no way to author a set of rules such that it can't be abused by someone with the power to enforce it. That's why the law has judges, juries, courts of appeal, standards of evidence, disclosure rules, corpus delicti, habeas corpus, and so on.

The problem is not the CoC, but the lack of accountability, oversight, and transparency in enforcement.

I totally agree, but people don't generally claim we should get rid of the law for these reasons. I am responding to the claim that we should get rid of CoCs, which I disagree with. I am not arguing that CoCs are perfect, or that we don't need checks. In this case, it seems that the CoC was pretty poorly written, and the team didn't even follow their own procedure. I think we should fix those problems rather than throw the whole thing out.

Agreed, but I don't see it happening. There's very little chance of any conference organizer setting up some sort of quasi-legal institutions around CoCs.

It is happening before our eyes! The organizers have admitted they made mistakes and are working on improvements. I agree a full-on court system would be ridiculous, but what that means is that the CoC needs to better thought out to reduce ambiguity.

Those are all valuable tools to have in a system where the law can deprive a person of property, liberty, and life.

It's not necessarily the case that you have to spend nearly as much on a code of conduct system where the ultimate worst-case scenario is that an innocent person is denied a platform or has the reputation tarnished in a reversible way.

I think you're saying that people who harass should be able to be removed absent a code of conduct. I agree. However, harassment isn't always the overt act you may be picturing, it can be more subtle. Code of Conducts give a single canonical state for what's considered acceptable, but more importantly, give organizers something tangible to point to when there is a violation. In addition to that, they inform people (even the harassed, who may not have considered what's happening to them harassment!) of what I'd acceptable.

As a white, male engineer, I haven't ever found a need for codes of conduct because I've had the privilege of never having been harassed. This isn't the case for all groups, so it's best to have one.

> Code of Conducts give a single canonical state for what's considered acceptable

This is what they're supposed to do. In reality, this is impossible because so many of the rules in a CoC are subjective, and therefore are enforced based on the opinions of the people enforcing the CoC.

> but more importantly, give organizers something tangible to point to when there is a violation.

This just gives the enforcers a sense of moral authority to impose their opinions, nothing more. I don't see that as a benefit of what is inevitably an incomplete document, and frequently poorly thought out as well.

> Code of Conducts give a single canonical state for what's considered acceptable

Except... they don't.

They say things like (quoting from the NumFOCUS CoC):

> We will not accept harassment or other exclusionary behaviors, such as:

> ...

> Other unethical or unprofessional conduct

That leaves things wide open -- whatever the committee decides is "unprofessional" is hereby banned. How am I supposed to guess their mind?

> If you need a CoC to get rid of someone who harasses others what you have is a deeper problem

How many conferences did you run before CoC's became popular?

Cue the cries of "I broke no rules!" when someone is "gotten rid of" without a CoC.

The task of making rules like that is never ending. This concept reminds me of when the head of GM rewrote their dress code. It went from several pages down to two words: "dress appropriately". What that means actually depends on context - who are you meeting with and what kind of work is going on. She realized that manager who needed a manual to figure that out had bigger problems.

I went to Cambridge University, UK from a working class background. Unwritten dress codes were one of the many things used to "other" anyone from the wrong background. These things aren't obvious.

Software developers, who are intimately familiar with the idea that software needs constant attention and maintenance, seem to often balk at the idea that communities also need constant attention and maintenance. The idea that a code of conduct is "never ending" does not seem like a bad thing to me.

As an example, the mozilla community participation guidelines (currently at version 3.1) are quite short and readable and many sections are essentially your "dress appropriately" example (i.e. "Be Respectful"), with quick a paragraph to clarify the idea.

https://www.mozilla.org/en-US/about/governance/policies/part...

>> The idea that a code of conduct is "never ending" does not seem like a bad thing to me.

So another area where that type of thing can happen is in union negotiations. Unions in the US started out fighting for basic decent treatment - reasonable hours and pay, safer working conditions, etc... Some years ago I was in a manufacturing plant and the break (lunch) room had a big television in it. Turns out the union had demanded the TV in their most recent contract. Someone I was speaking to berated the union "those idiots demand a TV, don't they have real concerns?" I realized the problem is that the adversarial nature had grown so bad - probably on both sides - that nothing could be had without a negotiation and putting it in writing. That goes for work from the union: "that's not my job" or worse - "you're getting written up for doing something that's someone else's job" to the management "no we're not giving you anything we don't have to by contract". Once you start writing things down and trying to nit-pick it can lead to a terrible place where nothing is easy for anyone.

They're meant as a reminder for people who actually read them, but otherwise they're meant as a list for enforcement. They're something to point to as "you knew the rules - that's why we're removing that person". That's why in this case the affected person complained about the CoC using very ambiguous rules that were up for personal interpretation.

And, indeed, the tech industry has had and continues to have a deeper problem.

If you need a system of written laws that apply to everyone what you have is a deeper problem

This is certainly true, but how else would you propose solving the deeper problem?

One option is to designate someone or some group as arbiters / benevolent dictators, and have the rule of "If so-and-so decides you're making the place worse, you're not welcome." It's certainly effective. But it exacerbates Jeremy Howard's complaint - which is not so much about CoCs per se as about the group of people who enforced them and the way in which they did so. I don't think getting rid of CoCs will really solve that problem.

One option is to have a closed or invitation-only group - but that's at odds with the goals of many communities. (And it doesn't reliably solve the problem, it just makes it less likely you'd run into it.)

Reply on news.ycombinator.com