Pick two: concise, comprehensive, or clear rules
post by Screwtape · 2025-02-03T06:39:05.815Z · LW · GW · 3 commentsContents
I. II. III. IV. V. None 3 comments
There once was a post on LessWrong asking what's with all the bans [LW(p) · GW(p)]? The OP of that post complained of, among other things, being punished despite not having broken any rules that they knew about before taking the action they got punished for. I started to write a comment, which then ballooned out of scope. The following is based on my own experiences trying to write rules.
I.
I claim the ideal rules are concise, comprehensive, and clear[1]. In practice, I think that we usually have to pick two.
Concise: Ideal rules are short. You can read or refer to them easily.
Comprehensive: Ideal rules include everything relevant. There aren't gaps.
Clear: Ideal rules have one interpretation. There's no ambiguity.
If you're running LessWrong, then "Don't be a jerk, talk about things related to rationality" is a concise and fairly comprehensive set of rules. It's just not a very clear set of rules. There will be myriad places where people disagree with what counts as being a jerk or what relates to rationality. "Do unto others as you would have them do unto you" is a famous concise and comprehensive rule with a lot of interpretations. If you aren't under a lot of heated or adversarial pressure though, concise and comprehensive rules are great! When they fail, it's because people don't agree on how to interpret the rules.
There's a subtype of concise and comprehensive that goes like this: "The moderators shall do whatever they want." This is very unclear! It is impressively comprehensive however, since your mods can do whatever they think is reasonable whenever an issue is brought to their attention. (Even easier if they don't need to be bound by precedent.) It doesn't scale especially well, since it's hard to automate mod judgement. Eventually your mods might disagree with each other, a problem which can be solved by the second rule "the moderators shall do whatever they want, unless overruled by the administrator."
The official rules for Magic: The Gathering are comprehensive and clear. Magic is played competitively by an often pedantic fanbase, the more competitive games are overseen by trained judges, and its rules are programmed into a computer which accepts no on-the-fly rulings. The comprehensive rules are also (at the time of this writing) two hundred and ninety three pages long and include blocks of text like the following:
109.5. The words “you” and “your” on an object refer to the object’s controller, its would-be controller (if a player is attempting to play, cast, or activate it), or its owner (if it has no controller). For a static ability, this is the current controller of the object it’s on. For an activated ability, this is the player who activated the ability. For a triggered ability, this is the controller of the object when the ability triggered, unless it’s a delayed triggered ability. To determine the controller of a delayed triggered ability, see rules 603.7d–f.
The majority of Magic players have never read the Comprehensive rules, because those rules are long and have a tendency to induce acute narcolepsy in everyone except modrons, rabbis, and NixOS developers. If you have to reference rules 603.7d-f, your ruleset no longer fits in the average person's head. Nevertheless, these rules can withstand a lot of pressure. When they fail, it's usually because not enough people are actually reading the rules. You also wind up with people standing just barely in or out of the line the rules, to the irritation of those around them. Consider U.S. tax law; most people need a trained helper or specialist software to figure it out. Even when the rules are followed exactly, you wind up with people angry at how much other people are or aren't paying.
Forum software which bans two or three slurs is clear and concise. If you type any of a specified list into the message box and hit submit, the banned word (or depending on the settings, the entire message) will fail to appear. This was a common setup on the internet bulletin boards of my youth, and caught a lot of very low effort spam or flaming.
As most of my peers experienced with internet bulletin boards are no doubt aware, circumventing a banned word list is laughably easy. If the mods just censor the word "idiot" then you can usually type "idi0t" or "id iot" or switch to synonyms like "moron" or "dumbo." I have played whack-a-mole with the creativity of the internet before, and while a banned word list might catch the very low effort posters it doesn't take much effort to get around it. When clear and concise rules fail, it often looks like someone exasperatedly following a diarrhetic dog around with a scooper. "No, you can't do that. Or that. Or that either. Why would you think this fourth thing was allowed? No, my shoes aren't better than the carpet." This process might gradually result in comprehensive rules, but by the time they get there the rules are no longer concise.
Clear and concise rules also tend to have really silly looking false positives. "Can I get some assistance?" contains the substring "ass" and therefore will get flagged by the simplest kinds of profanity filters. Quoth Patrick McKenzie, "I can safely assume that this dictionary of bad words contains no people’s names in it" is a falsehood many programmers believe.
II.
Step one is to have sympathy for the moderators and authors of such rules. The job is harder than it looks.
Step two is to notice the tradeoffs, and what situation you're in.
If the moderators are trying to handle a large population or a broad set of circumstances, the rules basically have to be comprehensive. That means you need to trade off being clear (allowing ambiguity and judgement call) and being concise (allowing the written rules to grow large and complex.)
If the judges are trying to let people make significant, long term investments based on the rules, then the rules basically have to be clear. Signing a mortgage agreement or life insurance payment on a quick conversation (concise) and broad statement (comprehensive) is inviting heated argument later. Citation: many angry divorce settlements.
If the setup relies on everyone understanding what's expected, the rules basically have to be concise. When I was getting my driver's license, I recall being given the rules of the road in a pamphlet. There weren't many pages in the pamphlet, because people who fail ninth grade English classes drive cars. People tried for years to write software that could drive a car (good software is by necessity comprehensive and clear) and failed; what seems to have worked is giant Machine Learning models which are not even a tiny bit concise.
Step three is to cry.
III.
This essay is an answer to people who are frustrated with where the rules are falling short.
When I was studying for my driver's license, I noticed that the pamphlet obviously didn't include all the rules of the road. I wanted to read the actual rules behind driving, and since I was on the cusp of adulthood I figured I'd read the laws of the country I lived under while I was at it. I was prepared for it to be hard, but I was a fast reader and precocious; it wasn't uncommon for me to read a textbook or a small stack of novels over a weekend. So I went to the library and asked for a copy of the laws of the United States of America please, yes, all of them.
This was when I found out that there are a lot of laws in the USA, and your local library or town hall didn't actually have a copy of the full text lying around available for high schoolers. I think I eventually talked my way into a college's law school library, where I found that even the people teaching lawyers and judges didn't have a copy organized for reading from start to finish. This made me pretty angry. If I was expected to live under the U.S. legal code, and would be punished if I broke the laws, then it seemed extremely unfair not to be given a list of the laws so I could know what it was I was supposed to obey. I didn't quite get all the way to radical anarchist rebellion, but I thought about it, and it's the seed of my empathy for anarchists these days. (I want to give a major round of applause to the age of the internet and whoever put together the Massachusetts General Laws web page. That's the kind of thing I was looking for as a teenager.)
If you are frustrated with where the rules are falling short, please take a moment, and ask yourself:
- Is your frustration that the rules are long and seeming too concerned with minutiae? If so, consider if it would be possible to make them much shorter without introducing ambiguity.
- Is your frustration that the rules don't cover an area of behavior you think is important? If so, consider if it would be possible to make them cover this and every equally important area without becoming overly long or using lossy summaries.
- Is your frustration that the rules are fuzzy or vague, relying too much on interpretation? If so, consider if it would be possible to make them unambiguous without taking a lot more words or ignoring some categories altogether.
There's an attempted synthesis, where comprehensive and clear rules have good summaries. I don't know all the legal definitions of theft are, but "thou shalt not steal" has worked pretty well for me my entire life so far. It's a good system! But it's a nontrivial abstraction, and sometime it leaks. All of these approaches leak, and when they do one of the three corners gets sacrificed.
- Sometimes the rulemaker says, look, I can't make this simple, you are going to have to read a lot of detailed rules and that's going to take a lot of effort. It is not concise.
- Sometimes the rulemaker says, look, I'm not trying to describe all of society and social behavior here, I'm not going to bother saying you can't stab your opponent in a chess game, you are going to have to assume these rules just cover a narrow case. It is not comprehensive.
- Sometimes the rulemaker says, look, people are going to disagree about the edge cases sometimes and I'm going to make a call as I see fit. I can't even say I'll make the calls in the same direction each time or eventually common law will grow too large. It is not clear.
Sorry. This is where we're at.
IV.
While I'm on the subject there's one more thing I'd like to note.
Some people behave as though nobody should be mad at their behavior as long as they're following the rules. This is not so. You can be deeply aggravating and follow the rules scrupulously. Consider the child who chants "I'm not touching you" in the backseat of a car, their finger held an inch away from their frustrated sibling. They might be obeying the rule "stop hitting each other" but we don't expect the sibling to feel happy about it.
"I followed all the rules [2] so why am I being punished by the arbiter?" is a fair question.
"I followed all the rules, so why are people mad at me?" is a mistake.
The arbiter is a person.
To pick a few examples: intentionally fouling the other players then accepting the penalty timeout won't make the people you fouled happy, constantly using the most offensive language that doesn't get you censored will lead some people to spend less time with you, saying things that are technically true but deeply misleading will still lead to some people trusting the meaning of your words less, and if you only play blue control decks I'm going to play Magic: The Gathering with somebody else.
(I have a Magic deck I built as a teaching tool. It contains nothing but counterspells, zero power creatures that blocked well, and hand attack spells. Its win condition was to mulligan to six, then wait for the other player to deck themselves via normal draws. Nobody liked playing against this deck, even though it was obviously legal and didn't even involve any weird obscure card interactions.)
Depending on the goals and priorities, sheltering people who are infuriating but not breaking any rules may be the right move. Twitter and Facebook are big spaces, and as long as there's useful blocking and feed filtering tools I'm inclined to defend infuriating people being able to use the platform. The local subway system is an important piece of civic infrastructure. As obnoxious as I find it when someone steps on board smelling of weed or wearing a shirt with foul language on it, I still think they should be allowed to use the subway like that.
For other goals, sheltering people who are infuriating but not breaking particular rules may be the wrong move. If you are sufficiently persistent about whatever it is you're doing that is not against the rules but is making people mad, people will either stop hanging out in the space you're in (including the person running the space stopping from doing that,) bend the existing rules such that it kind of covers what you're doing if we squint, or will make new rules that bans the thing you're doing. This is not unusual evidence of them being tyrannical or capricious. This is them encountering something they wish was different (your behavior) and trying to figure out how to get what they want.
I believe the best, most virtuous thing they can do in this circumstance is to say explicitly something like 'yep, we don't don't seem to be able to write the rules in a consistent way, but what you're doing is a problem. Stop it or leave.'
V.
I currently think LessWrong is trying to be concise and comprehensive, which means it relies on judgement calls by the administration and moderation team.
I think the higher the stakes are, the more we should aspire to be clear. If the government wants to put someone to death for breaking the rules, that rule should be damn specific.
I think that it's underappreciated the extent to which 'fixing' a rules system by pushing for one of these three risks sacrificing one of the others.
Finally, arbiters are in short supply. I appreciate them. All three virtues impose different costs on the arbiters.
- ^
This is an assertion. Read the rest of the essay to see if you agree with how I'm using those words.
- ^
Even the fuzzy ones or the ones from outside of the scope of this ruleset or the ones that aren't in the top five that everyone knows
3 comments
Comments sorted by top scores.
comment by Yair Halberstadt (yair-halberstadt) · 2025-02-03T09:22:09.394Z · LW(p) · GW(p)
This is all true, but sometimes people assume the opposite:
"Rules which are not-concise must be clear and comprehensive".
This is trivially false, but rules can often end up falling into this pit by keeping on adding new rules to handle unexpected edge cases. But each new rule creates further edge cases which have to be handled until the whole system becomes so complicated that nobody what's allowed or isn't and you call the work done.
Hence even in areas where comprehensiveness is important, like the tax code, it can be valuable to push for simplification. Because if verbosity isn't actually buying the comprehensiveness or clarity you need, might as well at least be concise.
comment by Knight Lee (Max Lee) · 2025-02-03T07:52:43.518Z · LW(p) · GW(p)
A lot of forums have open-ended rules which give moderators discretion (hence unclear), but in my opinion LessWrong takes the cake by not only having unclear rules, but unclear rules combined with high standards on a fuzzy "signal to noise" measure.
In this answer by habryka [LW · GW]:
Just because someone is genuinely trying to contribute to LessWrong, does not mean LessWrong is a good place for them. LessWrong has a particular culture, with particular standards and particular interests, and I think many people, even if they are genuinely trying, don't fit well within that culture and those standards.
[...]
Signal to Noise ratio is important
Thomas and Elizabeth pointed this out already, but just because someone's comments don't seem actively bad, doesn't mean I don't want to limit their ability to contribute. We do a lot of things on LW to improve the signal to noise ratio of content on the site, and one of those things is to reduce the amount of noise, even if the mean of what we remove looks not actively harmful.
I understand the motivation behind this, but there is little warning that this is how the forum works. There is no warning that trying to contribute in good faith isn't sufficient, and you may still end up partially banned (rate-limited) if they decide you are more noise than signal. Instead, people invest a lot only to discover this when it's too late.
I think there should be a clearer warning about this.
Let users decide what to see:
I suggest that instead of making rate-limited users (who used up their rate) unable to comment at all, their additional comments should be invisible, but still visible to other rate-limited users (and users who choose to see them).
Rate-limited users should see a special emphasis on comments by other rate-limited users, or normal users who choose to see invisible comments. This way they know who are able to read their comments and interact with them. The same applies to posts instead of comments.
I would like to see the comments by rate-limited users, and I think a lot of other users would want to see them. Anyone who once was rate-limited in the past would probably want to, and should be encouraged to.
Believe it or not, I haven't been rate-limited on LessWrong (yet!), but I've been banned from other places, hence this attitude.
comment by papetoast · 2025-02-03T07:22:42.384Z · LW(p) · GW(p)
As someone who wrote pages of pedantic rules for minecraft doors, I relate to this post a lot. Rules are just hard to write and to enforce consistently