Binary Translator

Why computers count in twos

Binary is not an arbitrary choice or a programmer's affectation; it falls out of physics. A computer is built from transistors, and a transistor is fundamentally a switch — it is either conducting or it is not. Building a reliable device that distinguishes two states is straightforward; building one that reliably distinguishes ten different voltage levels, and keeps doing so as components age and heat up and drift, is a nightmare. So the machine gets two symbols, 0 and 1, and everything else is constructed on top of that foundation. Numbers, text, images, audio, video and the programs that manipulate them are all ultimately long strings of those two states. Binary is not how computers represent data as a design decision — it is the only alphabet the hardware can actually spell in.

From ASCII to Unicode

The bridge from binary numbers to readable text is a character encoding: an agreed table saying which number means which character. ASCII, standardised in 1963, covered 128 characters — enough for English letters, digits, punctuation and a handful of control codes — and for decades that was the whole world as far as American computing was concerned. It was, of course, hopelessly inadequate for anyone who needed an accented letter, let alone Greek, Arabic, Hebrew, Devanagari or Chinese, and the result was decades of incompatible national encodings and mangled text. Unicode was created to fix that permanently by assigning a unique number to every character in every writing system humanity has used. UTF-8 is the encoding that stores those numbers efficiently, using a single byte for the old ASCII range — so plain English text is byte-for-byte identical to what it was in 1963 — and expanding to several bytes for everything else.