Surely you can see the point Oscar is trying to make, though? If a post's claim to topicality was that it presented "a rational way to do something", and if (by the advice of the original post) you instead make it present "a way to do something", then that raises the question of what it's doing here instead of on the rest of the internet.

I think my claim as stated is literally true, since if I could add that constraint as a community norm, I would.

I come to Less Wrong to learn about how to think and how to act effectively. I care about general algorithms that are useful for many problems, like "Hold off on proposing solutions" or "Habits are ingrained faster when you pay concious attention to your thoughts when you perform the action". These posts have very high value to me because they improve my effectiveness across a wide range of areas.

Another such technique is "Dissolving the question". The post you linked to (Yvain's "Diseased thinking: dissolving questions about disease") is valuable as an exemplary performance of this technique. It adds to Eliezer's description of question-dissolving by giving a demonstration of its use on a real question. It's main value comes from this, anything I learnt about disease whilst reading it is just a bonus.

To quote badger in the recent thread "Rational Toothpaste: A Case Study"

I claim a post on "rational toothpaste buying" could be on-topic and useful, if correctly written to illustrate determining goals, assessing tradeoffs, and implementing the final conclusions. A post detailing the pros and cons of various toothpaste brands is for a dentistry or personal hygiene forum; a post about algorithms for how to determine the best brands or whether to do so at all is for a rationality forum.

But we don't need more than one or two such examples! Yvain's post about question-dissolving was the only such post I ever need to read.

Posts about toothpaste or house-buying or room-decoration or fashion (EDIT: shaving, computer hardware) only tell me about that particular thing. As good as many of them are they'll never be as useful as a post that teaches me a general method of thought applicable on many problems. And if I want to know about some particular topic I'll just look it up on Google, or go to a library.

It's not possible for LessWrong to give a rational treatment of every subject. There are just too many of them. Even if we did I wouldn't be able to carry all that info around in my head. That's why I need to learn general algorithms for producing rational decisions.

(Also: Even though badger makes it clear in the quote I gave that the post is supposed to about the algorithms used, only one of the comments on it (kilobug's) is talking about this. Most of the rest are actually talking about toothpaste.)

EDIT: Should I repost this to discussion level?