Caret

The caret is a position, not a thing. It sits between two characters, and every editing action is defined against it: typing inserts there, Backspace removes what is behind it, Home and End take it to the ends of the line, and a selection is what you get when it is dragged rather than placed. Because it marks a gap rather than a character, it is drawn as a thin bar between letters, and the classic block cursor of a terminal is the same idea rendered for a grid where a gap has no room of its own.

Two words collide here and it is worth keeping them apart. The mouse pointer is the cursor in everyday speech and in CSS, where cursor: text names the I-beam shape the pointer takes over editable text. The caret is the insertion point, and it stays where it was put when the pointer wanders off. Apple’s guidelines say insertion point for exactly this reason. Two other senses of the word are near enough to trip over: the caret that means a dropdown’s small triangle (see disclosure triangle), and the typographic caret, the proofreader’s mark for inserted text, which is where the name came from.

It blinks because a stationary hairline is easy to lose and blinking is what draws the eye back to it. Around a second is the traditional rate, and it is one of the few animations that runs forever without asking, which makes it a fair thing to switch off: a stated motion preference should stop it, and it should never be the only thing on screen that moves. On the web, caret-color styles it, including transparent for interfaces that draw their own, which is the usual move in a code editor or a custom text engine.

Placing it is harder than it looks. A click has to be resolved to the nearest gap between characters rather than to the character under the pointer, which is why a press slightly right of a letter’s middle lands after it. Touch adds a magnifier, because a fingertip covers the very text it is aiming at. And a caret can exist outside a text field: caret browsing (F7 in several browsers) puts one in the page itself so a reader can move through the document with the keyboard, which is one reason content should survive being selected and traversed.