Comment by dang
3 years ago
Good point. The intro to the quiz asks you to answer the questions literally, but by asking this the author kind of assumes their own conclusion. I wonder how much consensus there would be if the intro asked you to go by the intent of the rule as you understand it, rather than what it literally says.
The first time I went through the quiz, I followed the instructions and had to think about definitions a lot. Then I read your comment and went through the quiz again and just used common sense (dangerous phrase, but I believe it worked in all 27 cases). There was only one violation: someone drove a Honda Civic through the park. What was that person thinking!
On HN we've always tried to avoid hair-splitting arguments by appealing to general values rather than trying to nail down the precise list of disallowed behaviors [1, 2]. Trying to be precise seems like a ticket to bureaucratic, soul-destroying hell [3]. I'd rather just say that there aren't precise rules, just an intended spirit and a few pointers, and yeah that means there's a lot of interpretation involved. There's going to be a lot of interpretation involved no matter what you do, so why pretend otherwise? Just make it clear up front. Then you can say "someone's got to interpret the rules, and that happens to be my job, and I'm interpreting them this way". People will get mad, but people are going to get mad no matter what you do, and at least you won't have to argue about whether a bicycle is a vehicle.
That doesn't mean there aren't edge cases and disputes about which calls are fair. There are tons of those. But if you don't try to be precise then at least you don't get into semantic hell. Except when you do. Boy this work's hard: https://news.ycombinator.com/item?id=11301437
> The intro to the quiz asks you to answer the questions literally, but by asking this the author kind of assumes their own conclusion.
...and yet in this very HN discussion we have large numbers of people disagreeing about the intent of the small, clearly written intro, with each side convinced that their interpretation is the obviously correct one. I feel this does as much to support the author's thesis as the game itself.
The whole discussion hinges on the definition of "vehicle", which is not defined. If I take my definition as being "a vehicle that meets the definitions of a motor vehicle in the UK Road Traffic Act 1988, but not including those in use as emergency service vehicles" then the majority are indeed correct. And that's how I answered, because it's consistent with my expectations for a sign like that which has been placed by a human authority. If you choose the rather broader definition (from Wikipedia) of "machine that transports people or cargo" then more scenarios are excluded. But that's a more fuzzy definition too: is a remote controlled car (neither carrying people nor cargo) really not included in the category of "vehicle"? In practice you won't typically find humans placing signs with that intent.
So I agree with the parent and grand-parent comment. Without a context in which to understand the terms presented, even an apparently-clear statement and introduction can be entirely unclear. I suspect that if you pointed to an actual physical sign and a scenario unfolding in front of them, and asked people whether they objected, rather than asking as an abstract concept on a computer screen, you'd get a different distribution of answers.
Right, but Dan only has to agree with himself, not with everybody else who reads the guidelines and tries to extrapolate from them.
> People will get mad, but people are going to get mad no matter what you do, and at least you won't have to argue about whether a bicycle is a vehicle.
The quiz epilogue said something along the same lines. Basically the point was to prove with these questions that corner cases always exist, and the system can never be perfect, and therefore we’re screwed and might want to give up. “pinning down a definition is usually impossible” … “You might think you can add enough epicycles to your rules to avoid this problem.” … “Maybe you will decide to live with the nebulosity, but have more sympathy for the refs. Maybe you will decide that you would prefer to live with the consequences of less moderation. Maybe you will think really hard about decentralization (which is not a panacea). Maybe you will give up on social media altogether.”
I do have sympathy for the refs, Dan, and I think you do an amazing job at a Sisyphean task. I’m also okay with nebulosity too.
However - I want to push back a little on the idea that we can’t or shouldn’t try to be precise, at least not as the most significant summary bit. We should try to be precise when we can, and provide examples when we can’t. I don’t buy the author’s argument/implication that the existence of a corner case somewhere means we shouldn’t be attempting to define the “epicycles” of the rules, especially when it’s really easy to say something like the park boundary is 200m above the ground, or insert ‘motorized’ in front of vehicles, which immediately eliminates like 50% of the supposedly hard to answer questions. Include the other rules, and add details to the quiz questions and almost all of them can become unambiguous. The point of all this is to provide clarity whenever possible and minimize the corner cases and reduce the number of people getting mad, right? It matters whether it’s just one or two people flaming each other versus everyone. It matter whether there’s only one or two crazy accidents in parks versus thousands or millions.
There’s a real difference between public safety and online forum opinions, of course. Yes, with a Grand-Canyon-sized gray area in between. But whether an airplane can fly through a park probably deserves a lot more bureaucratic attention than nailing down how people talk about Pi and religion on HN? Maybe I’m conflating law and forum moderation, maybe you were only talking about forum moderation, but I’m thinking about law as social moderation and how the quiz should reflect on social moderation in general. Our laws currently are in the process of building a larger and larger decision tree of both vague and specific language about what activities and behaviors are socially and legally acceptable, trying eternally to be more precise, and for the most part it “works” by some definition to keep the system manageable. We do try to get precise with speed limits and what kinds of death deserve what punishment and what constitutes insider information and whether badly compressed mp3s constitute copies. Even when it’s hard to pin down, we keep on trying, in order to reduce mistakes.
It’s kinda fun this little quiz of ambiguous questions caused so much discussion. Maybe it happened primarily because of the ambiguity, so each one is a little bike shed. Clearly the author said answer literally and most people just didn’t. But I somewhat feel like (maybe to the top comment’s point) that the contrived ambiguity backfired a bit on me. The problem with the quiz is withholding context and details in order to argue that it’s hard to draw lines. Context and details matter and they always exist in the real world. There isn’t only one rule, and a lot of the questions that seem ambiguous have actual right and wrong answers depending on details (e.g., altitude of the airplane & country of the park, or whether any country on earth asserts air & space rights hundreds of miles above their parks.)
That's a fairly culturally specific interpretation of common sense. Where I live it would for sure also include e-bikes and scooters, quite possibly regular bikes too (this is assuming "park" here means something on the ground and not e.g. a roof park where there might be weight limits).
The HN approach makes things simpler for moderators in much the same way that being a monarchy makes lawmaking simpler for the king, but writing down rules isn't about making the enforcer's life easier, it's about making the subject's lives easier. They're more numerous, so their needs should have at least some weight.
Independent of that argument, precise written rules and a process for updating them are valuable for several reasons:
1. Whilst people might still get mad, they get mad at the written rules and not at the interpreter of them. This takes a lot of the heat out of the situation because a document can be improved easily relative to improving a person, so discussions about bad outcomes become de-personalized and more constructive.
2. The act of writing down rules forces mental clarity. Contradictions and unhelpful biases that may not be obvious when free-floating in one's head can become apparent immediately when trying to write it all down.
3. Because the rules are clear, violations are less likely to happen to begin with. People who aren't on-board with the values of the community stay away.
The generic HN prohibition against "flamewars" is a good example of a rule that could use a rigorous clarification. It doesn't work to assume the intent or definitions are obvious, because flamewar is a purely online concept that doesn't have any clear analogy to the physical world. Actually it's the opposite: in physical debates there's a general understanding that anyone who turns up and takes part will engage in emotional self-control. If they lose it and start getting angry or raising their voice, they're the ones expected to leave, regardless of what argument the other side was making at the time. HN's approach inverts this standard social convention and blames the person who remains calm for the behavior of angry respondents!
The thread you linked to (from 2015!) is a good example of this. The original post is something about pi and the Bible. It's phrased calmly, isn't obviously in bad faith and is at least somewhat interesting yet is flagkilled, then you threaten to ban the user for conducting "religious flamewars". That user quite reasonably asks what it is that makes his post a rule violation and gives several possibilities e.g. is all discussion of religion banned? But you reply that it would be "soul destroying" to answer his question that specifically! He wasn't asking for a mechanical algorithm but getting more specific than "religious flamewar" and "spirit of the place" doesn't seem like an unreasonable request.
On the contrary, "precise rules" make things harder for many people in the community, because the more precise you get, the more attention people pay to them, and the more work they put into getting as far with those precise rules as they can (after all, if the rules are precise, surely it's OK to come right up to their edge, like the railings at a scenic overlook).
One weird subtext of this discussion is the idea that imprecision in the guidelines is costing "the subjects" something. But getting moderated doesn't cost you anything; on the contrary, it costs Dan. You just adjust and move on.
Getting moderated typically results in a post being made invisible to most site readers, which voids the effort put into writing it. Of course it imposes cost on the person being moderated, that's the purpose of moderation and bans in the first place.
Quite beyond the quiz instructions, the rest of the psychology on display in this thread is fascinating. Having clear rules makes things harder because other people will follow them. That sounds like the kind of outcome you fear only if you don't really want there to be rules, but only cultural homogeneity. Which is a stated anti-goal of the site.
2 replies →