Binary code translator

Type text on the left and binary appears on the right, eight digits to a byte. That covers both directions people ask for, English to binary going one way and binary to English coming back. Swap the boxes with the arrows; spaces and commas are ignored when decoding, so an unbroken run of digits works as well as neat groups.

Separator changes what sits between bytes in the output. Set it to none when you're pasting into code rather than into a message.

Every character has a number in the Unicode table. Write that number in base 2, pad it to eight digits, and you have a byte. Capital A is 65: 01000001. Lower-case a is 97: 01100001. The difference is 32, a single bit, and that was designed in so a machine could change case by flipping one wire.

Characters past the first 128 need more than one byte. UTF-8 spends two on most accented letters, three on most symbols and CJK characters, and four on an emoji. One of those produces 32 digits.

Binary arithmetic is far older than any computer. Gottfried Leibniz published it in 1703, and he wasn't chasing an application: he took a system where everything could be written with 1 and 0 as a demonstration of creation out of nothing.

The eight-bit byte is much newer and much more arbitrary. Early machines grouped bits in sixes and sevens. IBM's System/360 settled on eight in 1964 because it held two decimal digits neatly and left room for a bigger character set, and the industry followed. Nothing in nature requires it, which is worth remembering the next time a decode fails because the digits don't divide evenly.

Binary is the most verbose way to write bytes down. Base64 and hexadecimal carry the same bytes in far fewer characters, and they're what you meet in real files and headers.

For the number of each character rather than its bytes, use character codes or Unicode code points. And Morse builds a whole alphabet out of two states, the same way a byte does.