Paul Graham: How to write useful (on Habr)

(part 1)

What should be the essay? Many would say - convincing. They taught us that way ... But I think that we can strive for something more ambitious: the essay should be useful.

First of all, the essay should be correct. But just being correct is not enough. It is easy to make a statement correct by making it vague. This is a common flaw in academic writing. If you don’t know anything about the subject, you cannot be mistaken in saying that the problem is complex, that there are many factors to consider, that looking at it too simply would be a mistake, and so on.

Such statements are undoubtedly correct, but they do not tell the reader anything. A good text contains strong sayings - so strong that making them even a little stronger would mean turning them into a lie.

For example, it would be more beneficial to say that Pikes Peak is located near downtown Colorado than just somewhere in Colorado. But if I say that this is exactly in the center of Colorado, now it turns out that I exaggerate, because in fact it is a little east of the middle.

Accuracy and correctness are like opposing forces. It is easy to satisfy one, ignoring the other. Maintaining an academic style is bold but false, it is rhetoric and demagogy. Useful writing is bold and truthful.

There are also two other points here: telling people something important and something that at least some of them still don’t know.

To tell people what they don’t know does not always mean to surprise them. Sometimes this means telling them what they unconsciously knew, but never expressed through words. In fact, this can give more valuable insights, because it tends to be more fundamental.

Let's put it all together. A useful essay tells people something true and important, something that they did not know yet, and does it as unambiguously as possible.

Note that this is all a matter of level. For example, you cannot expect an idea to be new to everyone. Any insight that you had probably also had at least one of the 7 billion people living in the world. But actually enough if the idea is new to a large number of readers.

The same goes for correctness, importance, and directness. In essence, these four components are like numbers that you can multiply to get a utility score. I admit that this is a clumsy simplification, but nonetheless true.

How can you be sure that what you say is true, new and important? Believe it or not, there is a trick for this. I learned this from my friend Robert Morris, who is afraid to say stupid things. His trick is to say nothing until he is sure it is worth hearing. Therefore, it is difficult to get his opinion, but when it succeeds, it is usually true.

In the sense of creating an essay, this means that if you write a bad sentence, you do not publish it. You delete it and try again. Often you give up entire branches of the story from four to five paragraphs. Sometimes - from a whole essay.

It is impossible to guarantee that every idea that comes to mind is good, but you can make sure that every published one is good. Just not publishing those that are unsuccessful.

In science, this is called bias and is considered a hindrance. When some hypothesis leads to inconclusive results, in any case it is supposed to convey them to people. But when writing an essay, selection is a natural way.

My strategy is to loosen and then tighten. I am writing my first draft essay quickly, trying all the ideas. And then I spend a long time rewriting it carefully.

I never tried to count how many times I proofread an essay, but I'm sure there are sentences that I read 100 times before publishing them. When I correct an essay, there are usually passages that look annoying, sometimes because they are awkwardly written, and sometimes because I'm not sure that they are true. I get annoyed unconsciously, but after the tenth reading or so I say "Ugh, this piece" every time I catch it. They become like bulls that bite your sleeve when you pass by. I usually don’t publish essays until they all disappear - until I can read everything without any burrs.

Sometimes I skip a sentence that seems awkward if I can't figure out how to rephrase it, but I will never consciously miss a sentence that seems wrong. And you should never miss this. If the suggestion seems wrong, all you have to do is ask why this is not the case, and usually a replacement occurs right in your head.

To be continued ...

Who else wants to help with the translation, join us (Googledock shared for editing).