The Best Books I Read in 2021

One sentence from the most technical book on this list ended up explaining the other five. It is the First Law of Software Architecture, from Richards and Ford, and it goes like this. Everything in software architecture is a trade-off.

And that is a sentence about software. And then I spent a year finding it everywhere else.

Obama priced a decision. Gates priced a planet. Gawande priced a skipped step. Gladwell priced the luck nobody wants to count. García priced a day. Every one of the six put a number, or a cost, on something people usually argue about with adjectives.

So that’s the frame. Six books, one law.

This was my heaviest reading year and easily my most five-star one, and cutting it to six was genuinely hard. Here’s what survived.

A Promised Land by Barack Obama

Seven hundred pages, and it only covers the first term up to the bin Laden raid. There is a second volume still coming, which tells you what level of detail he is working at.

And what struck me was the doubt, which is a strange thing to find in a presidential memoir. He walks you through the auto bailout, the healthcare fight, the Afghanistan surge, and every one of them arrives with the version where he chose the other way, laid out with enough care that you can actually argue for it.

So what does that cost him? Certainty, which is the only currency a memoir like this normally spends. And most books in the genre are written to make a legacy look inevitable, and this one keeps showing you the fork.

Now, it is still an Obama book, and he is a careful writer about himself. So is some of it self-serving? In places, obviously, and any memoir written this close to the events is partly a brief for the defence. But the deliberation is real, and reading it slowed me down in meetings for a while afterward, in a way I would like to say lasted.

That’s the first price. Deliberation is expensive at the time and cheap in hindsight, which is exactly backwards from how it feels.

How to Avoid a Climate Disaster by Bill Gates

Gates opens with two numbers and never lets go of them. Fifty-one billion tons of greenhouse gases a year. Zero, which is where it has to end up.

And then he does the thing almost nobody does in this conversation. He breaks the fifty-one billion into sectors and puts a share on each one. Making things, cement and steel and plastic, is about thirty-one percent. Plugging in is twenty-seven. Growing things is nineteen. Getting around is sixteen. Keeping warm and cool is seven.

So why does that breakdown matter so much? Because it kills the argument about which single thing to blame. Transport is the sector everyone pictures, and it is the fourth largest item on the list.

And his other useful invention is the green premium, which is just the extra cost of doing something the clean way. Name that number per sector and you have a to-do list instead of a mood.

And the cleanest example in the book is jet fuel. Conventional jet fuel runs about two dollars twenty a gallon. Advanced biofuel runs more than five. That gap is the entire policy problem, stated as a number somebody could actually go and close.

That’s the second price. Urgency will get you attention, but arithmetic will get you a plan.

The Checklist Manifesto by Atul Gawande

Gawande’s argument is almost insulting in how simple it is. Experts skip steps under pressure. A checklist catches the skipped step.

And the idea comes out of aviation, from a 1935 crash. Boeing’s Model 299 went up, stalled and killed the test pilot, and the plane was written off in the press as too much aircraft for one man to fly. The pilots disagreed. They wrote a checklist, and that aircraft went on to fly as the B-17.

And he has the numbers. Peter Pronovost put a five-item checklist on central line insertions at Johns Hopkins, covering things every doctor already knew, like washing hands and draping the patient. The ten-day line infection rate went from eleven percent to zero. Then the WHO surgical safety checklist ran across eight hospitals around the world and cut major complications by roughly a third and deaths by nearly half.

Five items, one of them hand washing. That is the entire intervention.

And I run production data systems for a living, so this one landed on me differently. Most outages I have seen came down to somebody skipping a step that everybody in the room already knew, at two in the morning, because this one looked simple. The missing skill is almost never the problem.

So what does a checklist actually cost? Ninety seconds and your ego. That is the entire bill, and it is the reason people skip it.

That’s the third price. The step you skip is always the one you knew best.

Outliers by Malcolm Gladwell

Gladwell opens with Canadian junior hockey rosters and a birthday problem. A wildly disproportionate share of elite players are born in January, February and March. So why would a birth month decide a hockey career? Because the age cutoff falls on the first of January. And at nine years old, a boy born that month is nearly a year bigger than one born in December, right when the first selections get made.

That is the whole mechanism. Better coaching, more ice time, more games, compounding for a decade off a birth certificate.

And the ten thousand hours idea from this book has been argued about ever since, fairly so. Anders Ericsson’s original research was more careful than the version that went viral. So take the shape and hold the number loosely.

So what is the actual price here? Comfort. It is a nicer story to believe the best person won on merit, and Gladwell keeps handing you the upstream advantage that made the merit possible.

And he runs the same trick on Bill Gates, who happened to get access to a computer terminal at his school in 1968, when almost no teenager anywhere on earth had one. Gates put in the hours. Somebody else put in the terminal.

That’s the fourth price. Talent explains who got good, and circumstance explains who got the chance to.

Fundamentals of Software Architecture by Mark Richards and Neal Ford

This is the one genuinely deep technical book on the list, and it gave me the sentence I opened with. Everything in software architecture is a trade-off. That is their First Law, stated plainly and then defended for four hundred pages.

And their Second Law is the one I quote more often in reviews. Why is more important than how.

And here’s why that matters on a Tuesday. Every architecture decision I make buys something and sells something. Microservices buy independent deployment and sell you a distributed systems problem you did not have. An event log buys replay and sells you storage plus a schema you now have to version forever.

So what does the book refuse to give you? A right answer. But there is a set of trade-offs your team can live with. The work is finding out which set, which is slower and much less satisfying than reaching for a best practice.

That’s the fifth price. A best practice is somebody else’s trade-off, made in a room you were never in.

Ikigai by Héctor García

I closed the year on this one deliberately, after five books about decisions, costs and hard trade-offs.

So García and Miralles went to Ogimi, a village in Okinawa with an unusual number of people past a hundred, and asked them how they live. And the answers are unglamorous. Moai, which are small lifelong groups that check on each other. Hara hachi bu, which is eating until you are about eighty percent full. A reason to get out of bed that is specific enough to say out loud to another person.

Now, one honest note. The four-circle diagram everybody shares, the one about what you love and what the world will pay for, is a Western overlay rather than the Japanese idea. But the book is much more interested in the small daily business of a long life.

So what does it price? A day. And by the way, that is the hardest unit to think in, because a day is too small to feel like it is deciding anything.

That’s the sixth price. A life is bought a day at a time, at a price you barely notice paying.

And here’s where I should be honest about the frame. I found the trade-off pattern after the fact, because I happened to read the software book in the middle of the year and its sentence stuck to everything I picked up afterward. Read a different book in June and I would be telling you a different story about the same shelf.

That’s my 2021 list, and it is probably the hardest one I have ever had to cut down to size.

A best practice tells you what worked somewhere else. A trade-off tells you what you are about to give up.