I think your points about why courtesy is better are missing a crucial point; which is that social interaction is all about standardisation.

Consider the old VHS versus Betamax problem (or Blu-ray versus HDD for the modern version): two systems that achieved more or less the same goals, each of which had certain advantages and disadvantages going for it. But inevitably one system became popular and it stopped being economical for manufacturers to keep making both players and media of the other type, because not enough people would have used it. And this is a good thing because it means manufacturers don't have to produce media in both types, which means that the cost for the media that they do produce is slightly lower, and everyone except the die-hard users of the dead format wins.

Methods of social interaction are the same: you need both people who produce a certain kind of interaction and people who welcome those kinds of interactions. Regardless of which is better, the equilibrium point is towards one standard dominating - and the one that does dominate isn't necessarily better, it was just the first to gain critical mass.

That said, my intuition is that politeness is better than not-politeness in most contexts because it allows more plausible deniability. And that intuition is resting on the assumption that most people are highly protective of their status and therefore avoid status hits at all costs, even if that means taking twice as long to get to the point.

Describe a common situation where there is clearly more utility in stating x bluntly than stating x politely.

Here is a fairly narrow one: when you are correcting someone who has made a serious error which they will immediately recognize as an error when it is pointed out to them.

An example took place earlier here on this thread. Lionhearted had just stated that he would bow out of the discussion now. Wedrifid misread what was written, seeing "I'm bowing out for now", where lionhearted had actually written "I'm bowing out now". Wedrifid responded intemperately, making a particularly big deal of the withdrawal "for now", interpreting it as a kind of threat to return. (This comment has since been deleted by its author.)

I pointed out wedrifid's error bluntly, and was even so discourteous as to tease him on his embarrassing error. I am confident that this was the right way to handle this kind of mistake. Anything softer would have been condescending.

So that is one situation where bluntness strikes me as clearly best. But I'm not sure that this situation generalizes well. If the mistake were less serious (a typo, say) then the superiority of bluntness is debatable. If the mistake were less clearcut, then it would probably be wise to include some justification of the judgment that it really is a mistake.

Now that I've seen the issue framed in those terms, I can think of several cases where someone spent so long on niceness-padding that I got annoyed, lost interest, or interrupted to ask them to get to the point. I would like to add that the niceness/efficiency tradeoff is continuous, not discrete, bounded on the maximally-efficient end and unbounded on the maximally-nice end, and that there must be some amount of niceness-padding so excessive that will annoy even those who prefer prefer more of it in general.

If you're a mapmaker, being purely descriptive rather than prescriptive is the whole point.

Indeed, excellent counter-example. I was wrong to say there is no point in being descriptive.

If you mean that it's definitely more useful for people to behave in the way you prefer, you have not yet convinced me of that.

I am not sure that it is more useful. There appears, to me, to be some correlation between intelligence and blunt communication (nerds speak bluntly, mundanes politely) but that could be intelligence and contrarianism, or any other of many potential factors. I am not giving it any weight. However, I do think it's the case that it is useful for "people who behave in this way" to congregate and continue to behave in this way with each other.

That is, when their value systems are sufficiently similar in relevant areas, I can say that being more polite is an error for them. And LessWrong is one place where the value systems sufficiently coincide.

Well, okay, I think I see something. This post and this comment communicate that people often use arguments to support monolithic beliefs. People might be thinking that the first option you gave earlier is saying "I'm on your side, we need to work together to kill this 'unfeasible' enemy soldier" whereas the second is more like "Your idea is dumb and I will throw arguments at you until you retreat or surrender".

A kind of politeness, then, I could use, encourage, and appreciate when receiving would be an effort to communicate that we both want the outcome, your method has obstacles, can we fix these obstacles or find another route?

Thanks. It's an interesting example of balancing the need for clear rules without having so much specificity that it's easy to game.

I found this experiment with requiring non-redundant communication, but nothing about how long it went on or how it worked out in the long run.

It's the prescriptive/descriptive divide. When I say it's the wrong understanding, I mean that if I were to prescribe what understandings people ought to have of communication protocols, I would be in error if I prescribed this one. This understanding is worse than another understanding they could have. There doesn't seem to be any point being purely descriptive about anything.

You can play along and be accepted/effective in that culture, or not. It's your choice.

False dilemma. I can agitate for change in that culture.

It isn't dishonest to say "this is a good idea, but it might be difficult for these reasons" rather than "your idea isn't feasible for these reasons"

Part of my criteria for determining goodness of ideas is feasibility. I would be being dishonest if the idea was not feasible.