What I Want to Be Doing at Seventy

The textbooks that shaped me were written by Hennessy and Patterson — and decades later, at Google, I watched them still going.

A look back — long before any of the tools we argue about now.

My most vivid memory of the major starts in my second year, with Professor Soon-hoe Ha, newly arrived. He taught two required courses at once — Circuit Theory and Logic Design — so if you attended properly you sat in front of him at least six times a week. New material meant no old exam files to lean on, exams of extreme difficulty, and answers full of decimals that never divided cleanly.

In my third year I met him again, for a course called “Micom” — said to be different every year; whether it was “mycom” or “micom,” even what its real title was, my memory is hazy. But the textbook’s authors were, already then, god-tier: John L. Hennessy and David Patterson. For some reason those two won’t be remembered apart — living textbooks of computer architecture, Turing Award winners for 2017, and, years later, people I’d catch sight of from a distance at Google and quietly admire.

Computer Organization and Design

My memory had been distorted for a long time, so it took a while to sort out, but the Micom book was probably this one. I remember calling it “the abacus book”; I don’t recall ARM or MIPS in the title back then, and since the edition I found now has ARM, MIPS, and RISC all in it, real systems must have come in later. The Korean translation carries the professor’s name, so this is likely the one.

I remember taking the exam open-book, and the thing read so well for its size and depth that for a good while I labored under the delusion that my English was excellent. Getting to experience machine-translated hardware running up close was fresh, and a good memory.

Computer Architecture: A Quantitative Approach

I went to a computer-architecture lab for grad school and met this textbook in a graduate course — again by those masters. It carried not just the class but my whole life at the time. A book that made me think this is how a textbook ought to be made, and also a book of love and resentment, since I never mastered it fully and it made grad school hard.

My advisor told us to memorize papers, and gave exams every day. When there was no suitable paper, he had us literally memorize the important chapters of this book, one paragraph a day, for a long stretch. There were more failures than successes and I struggled — but looking back, as an English textbook it was a wise teaching I only understood much later. Had I been writing English papers I’d have gotten harsher guidance; instead, memorizing an excellent textbook, I turned words over — the shape of a paragraph, the use of demonstrative pronouns, different words for the same meaning — and drilled all of it.

The Masters, at Google

Twenty-some years after graduating, hearing the masters had won the Turing Award, I found myself with chances to see them from a distance — people I sort of “knew” — and through the odd guest lecture I was struck all over again. The high point was part of the 2018 Google I/O. I’d been thinking I couldn’t look deeply into AI and ML for all the energy that building and running services took; watching this man’s path, I decided not to call it “too late,” and resolved to follow the deep-learning papers. I’d be glad to grow old like that.

Part of 90s Computer Science Stories — first-person notes on growing up with computers and studying CS in Korea in the 1980s and ’90s. See the full series →

Adapted from my Korean essay on Brunch: brunch.co.kr/@chaesang/48