He invented null, then apologised for forty years

Tony Hoare died on the tenth of March.

If you write software you use his work today. Quicksort is his. Hoare logic, the way we reason formally about whether a program is correct, is his. CSP, which is why Go has channels, is his.

And null is his too. He put null references into ALGOL W in 1965, and in 2009 he stood up at a conference and called it "my billion dollar mistake".

The real cost of null is almost certainly far higher.

The apology

Plenty of people have shipped a bad idea. Few have owned one the way he did.

He could have defended it as a product of its time, or pointed to the constraints of 1965 hardware, which were real and would have excused him. Instead he said it was easy to implement, so he implemented it, and it had caused innumerable errors, vulnerabilities and system crashes for forty years.

Then he kept saying it, for the rest of his life, in public, to rooms full of people who admired him.

He was one of the most decorated people in the field (a Turing Award, a knighthood), and he had every reason to let it go. Nobody was going to take the award back.

What a senior engineer is

We have a broken idea of seniority in this industry. We treat it as the point where you stop being wrong.

Hoare is the counter-example, and he is about as senior as the field gets. Seniority is the willingness to name your mistakes precisely, in public, long after you could have quietly moved on.

The junior engineer hides the mistake because they think it threatens their standing. The senior engineer publishes it because they understand that the mistake is the most useful thing they own, the one piece of knowledge nobody else can give the team.

For anyone building a team, this behaviour spreads by copying. People do it when they watch someone senior do it and survive.

The obituary is worth reading.

Then go find the last thing you got wrong and say it out loud to someone junior to you.