TM★

Typesetter lab

The layout primitive CSS doesn't have: lines split around the obstacle — text on the left of the picture and on its right, same baseline. Every line here is placed by hand in JS: words measured with canvas, columns walked in line-height steps, the image subtracted from each band. Drag the picture anywhere; the type re-pours around it. Two composers: greedy (first-fit, ragged — what browsers do) and Knuth–Plass (TeX's algorithm — picks the whole paragraph's breaks at once and justifies the lines).

img 1 @ col 1 · 120px down

The text you're reading isn't really text — not the way a browser means it. Every line here was measured, broken, and placed by a small typesetting engine running inside this page, one word-width at a time.

It exists because the web can't do this. CSS can push words to one side of a picture, but it cannot split a line around one — ask a browser for a magazine and it hands you a scroll. Drag the picture and watch the paragraphs pour themselves again. That's not a video. That's arithmetic.

I build things like this for a living, which is a strange sentence, because for a living I have also mapped mines in 3D, published machine learning that predicts injuries before they happen, taught courses for Google, Meta, and Microsoft, and run a working farm on software I wrote between harvests.

The pattern underneath is consistent: someone hands me a problem nobody has scoped — a mine that can't see its own trucks, a course that doesn't exist yet, a farm that can't sell its own vegetables — and I figure out what to build. Then I build it.

Sixteen years of that teaches you the shape of unsolved problems. They are almost never unsolved because they're hard. They're unsolved because they sit in the gap between two crafts, and everyone stands politely on their own side, waiting.

This page sits in the gap between typography and code. The last people who stood here in earnest worked at Adobe, a decade ago, and moved on. The hard part was solved by Donald Knuth in 1981 — someone just had to walk over and pick it up.

So this is a portfolio, technically. Mostly it's an argument: the interesting work lives between the fields, and I've made a career of walking into the space between them. Staring into a gap like that? ⟦Say hi at tylermorgan.me/contact⟧

Either way — thanks for reading the columns. The picture's still draggable.

⚗ Back to the lab