Kerning by hand, and when not to bother

14 August · composition

Digital kerning is an opinion you can revise. Hand kerning is a piece of copper, and once it is in the stick the line is that much longer for good. The difference changes what is worth doing.

Where it pays

Display sizes, and only display sizes. At 24 pt a bad Ta pair is visible across a room, and one copper thin removed from the wrong side fixes it in about forty seconds. At 10 pt nobody has ever noticed, including the person who set it, and the time is better spent on the makeready.

The other case is capitals in a title. Foundry capitals were fitted for use in text, mixed with lowercase, which means a run of them set solid looks tight and lumpy at the same time. Letterspacing a titling line with 1 pt brass throughout, then pulling individual pairs back with copper, is worth the half hour.

Where it is vanity

Any body text. Any line already fighting the measure. Any pair involving a comma, because the comma is doing structural work and moving it reads as a mistake rather than as refinement.

If the correction takes longer than resetting the line, the line was wrong, not the fit.

What actually goes wrong

Kerning a line and then forgetting it during justification. The copper is thinner than a hair space and disappears in the stick; the line then justifies short by exactly the amount removed, and the whole paragraph creeps during lockup. The fix is procedural, not typographic: kern first, justify second, and count the line twice before it goes to the galley.

The forme in the front page photograph failed for a version of this. Two pairs were pulled in the second column after justification, the column went two picas short, and the furniture that should have caught it was the wrong length to begin with.