Preface

For about fifteen years I have been saying the same dozen things to anybody who worked with me.

They came out in meetings, mostly, as the short form of an argument I did not have time to make. One cannot act upon that which has not been defined. Do the same thing the same way every time and you'll get a consistent result. If it lives in your head, it should live in the system. People started repeating them back to me, which was gratifying and should have been a warning, because a sentence that gets repeated is a sentence that has stopped being examined.

I never wrote any of them down in one place. I never checked whether anybody serious had thought about them before, which is an embarrassing thing to admit about beliefs I had been operating on for a decade and a half. And I had never once gone looking for the case against.

This book is that, done late.

What prompted it was ordinary. Somebody in a meeting repeated one of my sentences back to me, approvingly, as the end of a discussion, and I realized I could not have defended it if they had pushed. Not because it was wrong. Because I had never had to. A rule that everyone around you accepts is a rule that stops being argued with, and a rule that stops being argued with stops being a rule and becomes a piece of furniture.

The version of that which should worry anybody is the one where the sentence is doing real work. These twelve have governed how I build things, which means they have governed what got built, which means the gaps in them are sitting in production somewhere with my name on them.

What it is: twelve rules I actually use, stated exactly as I have been saying them, then taken apart.

For each one I went and found where the idea came from, because none of these are original and pretending otherwise would be the first thing to disqualify the book. Most of them have a serious intellectual ancestry, some of it a century old, and in several cases the people who got there first stated the rule more carefully than I do. Then I looked for what the evidence says, including the evidence that contradicts me, and there is more of that than I expected. Then I went looking for the single strongest argument against each rule and wrote it as persuasively as I could manage before answering it.

Eleven of the twelve came out narrower. Two new ones went in. One had its central promise withdrawn and replaced with a smaller claim that happens to be true.

What it is not: a memoir, a management method, or a claim of originality.

Not a memoir, because the personal material in here earns its place by being evidence rather than by being mine. There is a morning routine, a commute, a weekend task list, and a drive-thru on Route 8, and each one is in the book because it demonstrates something a hypothetical example could not. Nothing has been invented to make a point land better.

Not a management method, because a method is a thing you can hand somebody, and this is a set of beliefs with their limits marked. The closest it comes is a chapter near the end that turns everything into seven questions, and even that is a conversation rather than a form.

And not a claim of originality. Almost every idea in here belongs to somebody else. What is mine is the particular combination, the ordering, and the specific way each rule fails, which I know because I have been failing that way in public for years.

One commitment, stated here so it can be held against the rest.

Where a counter-argument beat one of my rules, the book says so and the rule changes. That is not a rhetorical posture. Several chapters end in a worse position for me than they started, one of my favorite rules turns out to rest on a literature whose central model is not empirically supported, and the rule I use most confidently is the one the evidence treats most harshly.

I have tried to write the opposing case well enough that a reader might put the book down agreeing with it. That seemed like the only way to make the exercise worth doing. A creed that has never been at risk is not a creed. It is a set of preferences with a vocabulary.

A word on who this is for, since the answer shaped how it is written.

It is for me first. That is not modesty and it is not a disclaimer; it is a fact about the construction. I wrote it to find out whether things I had been saying for fifteen years would survive being checked, and the audience for that answer is the person who has to keep saying them. Several chapters are uncomfortable in ways that would have been easy to soften for a general reader and are not softened, because softening them would have defeated the purpose.

If it is useful to anybody else, the people it is likely to be useful to are the ones who have their own version of this list. Most people who have worked at something for a long time do, whether or not they have ever said it out loud. The specific rules probably differ. What almost certainly does not differ is the shape of the problem: a set of beliefs built from real experience, stated in a form that claims more than the experience supports, never tested against the cases they would fail, and getting quoted back at you by people who assume you have done the work.

There is one more audience I want to name because the book is partly addressed to them by accident. Everything in here is a method for building systems that decide on people's behalf, and it is written from the chair of the person who builds rather than the person decided about. That gap is the single largest thing I found and did not know I was looking for, and two of the fourteen rules at the end exist only because of it.

A practical note. The chapters are meant to be read in order, because the four families depend on each other and Part II reconciles against what Part I actually argued rather than against what it set out to argue. But each chapter also stands alone, ends by stating what it concluded, and says plainly where it gave ground. If you only want the result, Chapter 19 has all fourteen rules with their originals next to them, and it will make sense without the rest. It will also be considerably less convincing, which is the correct relationship between a conclusion and the work behind it.

The title is one of the twelve, and it is the one that gets quoted back to me most often. If it lives in your head, it should live in the system.

I meant it as an instruction about record-keeping. Writing this book taught me it is a claim about what can be moved between a person and a structure, that the claim is false in its strong form, and that what actually survives the move is smaller and stranger than I thought and still worth the effort.

That is roughly what happened to all of them.