The Printer's First Unkerned Type: On the Invisible Gaps of a Perfectly Set, Yet Unreadable, Page

Long before the first <div> was ever floated, the craft of arranging letters for a reader was a physical, painstaking art. In the print shops of the 15th century, typesetters would carefully place individual metal sorts—each one a letter, a space, or a piece of punctuation—into a composing stick. The goal was a perfectly justified block of text, a river of black ink flowing evenly down a page of pristine paper. But for the early printers, a peculiar and frustrating problem often emerged in their first proofs: certain combinations of letters would appear to have a jarring, awkward gap between them.

An uppercase ‘A’ next to an uppercase ‘V’ would sit too far apart, creating a visual hole that disrupted the rhythm of the word. A lowercase ‘y’ followed by an apostrophe might collide. The individual characters were perfectly formed, the leading (the space between lines) was consistent, and the block was technically aligned. Yet, the page felt amateurish and, more importantly, was difficult to read. The text was present, but the experience was broken. The letters were not in communion with each other.

The Space Between the Sorts

This was a problem of kerning, though the term itself came later. The word derives from the French *corners*, meaning “projecting angle,” and originally referred to the part of a metal sort that overhung its body, allowing it to sit over an adjacent character. Early typefaces, however, were often cast on uniform, rectangular bodies. Each sort lived in its own self-contained block, its bounding box utterly inviolable. A ‘T’ and an ‘.’ had to occupy the same physical width on the type block, leading to inevitable and unsightly gaps.

The solution wasn't to redesign the entire font but to adjust the relationship between specific pairs. Master printers learned to manually file down the metal blocks of troublesome letters or craft special sorts with overhangs. They were, in essence, creating the first bespoke spacing metrics. They were writing a crude, physical CSS for a handful of critical pairs: `letter-spacing: -0.5px;` but with a metal file. They understood that the perceived space—the optical geometry—mattered far more than the mechanical reality of the type block.

In our world, we face a digital echo of this ancient typesetter’s frustration. We load a beautiful, custom web font via a modern CDN. The file is optimally compressed, the `font-display` swap is set, and the Core Web Vitals report a blazing-fast LCP. By every quantitative measure, the page is a performance triumph. But for a critical moment, as the fallback system font swaps out for the intended typeface, the text violently reflows. Headlines break onto new lines, body text shifts down the page, and the user loses their place. The individual assets are optimized, but the relationship between them is unstable.

We have become masters of the digital sort, expertly optimizing each component in isolation. But like the early printer, our craft is judged not by the perfection of the pieces, but by the harmony of the whole. It is the invisible, negotiated space between the letters—the `font-feature-settings`, the `font-variation-settings`, the careful management of the FOIT and FOUT—that transforms a mere collection of characters into a seamless, stable, and truly readable experience. The goal is no longer just to render the text, but to honor the space between.

Notes & further reading

A few pages I came back to while writing this: