REEL 01 · KTM → USA

the story

Most portfolios hand you the finished print. This one runs the negatives instead: five frames, in order, starting with a machine I broke.

the machine that kept breaking
every crash was a puzzle
turns out it has a name
reliability under uncertainty
the next frame
FIRST BOOT · WIN98
frame 1 / 5mid 2000s
ch. 01mid 2000s

the machine that kept breaking

I started on a Windows 98 machine that I kept breaking. It was the family computer, and I was the reason it needed a reinstall every few months. Blue screens, registry rot, mystery slowdowns that always showed up the week after I had been somewhere I should not have been. I caused most of them and wanted to understand all of them.

the machine that kept breaking
FIRST BOOT · WIN98
ch. 02late 2000s

every crash was a puzzle

Every crash turned into the same question: what did I change, and how did that end up here? At that age I almost never got an answer I liked, and I kept asking anyway. Reinstalling Windows became something I could do half asleep. The machine was far more interesting broken than it ever was working.

every crash was a puzzle
STATE · UNKNOWN
ch. 03undergrad

turns out it has a name

Somebody had already written a whole degree for people who break things and need to know why. Four years of it, and the parts that stuck were rarely the parts being examined. Lectures handed me vocabulary for what I had been doing since I was a kid. Group projects taught me that systems fail at the seam between two people who each assumed the other had it covered. And every famous outage we picked apart looked like my Windows 98 machine with more zeros on the damage. That is roughly when I stopped treating failure as bad luck. It is almost never one mistake. It is one mistake that got company.

turns out it has a name
ROOT CAUSE · PLURAL
ch. 042022 — now

reliability under uncertainty

Fifteen years later I'm doing the same thing on a payroll, on distributed systems serving hundreds of thousands of people. The failures are subtler now and the pager is much louder. Nobody lets me reinstall anything, which I miss more than I expected. The questions got bigger, and I'm not convinced the curiosity behind them changed at all.

reliability under uncertainty
UPTIME · 99.9
ch. 05next

the next frame

There's a depth to this that a full-time job doesn't leave room for, so the next stretch is going back to study, most likely abroad. What I keep circling is keeping a person in the loop: systems someone can still open up and argue with. The kid who broke the family computer would like to make a career out of asking why things fail. Ideally before they do.

the next frame
KTM → USA
( 02 )

off the clock

click to develop