What Is Unicode?

Early computers used ASCII, a standard from 1963 that set aside just 128 numbers for text — enough for the English alphabet, digits, and basic punctuation, but nothing else. That worked fine for English, but was useless for languages with accented letters, entirely different alphabets, or thousands of characters like Chinese.

So every region invented its own patchwork fix: dozens of separate, mutually incompatible "code pages," each reusing the same numbers to mean completely different characters depending on which one a computer happened to be using. Open a file created with the wrong code page and you'd get mojibake — rows of garbled symbols instead of the intended text — because the number 233 might mean "é" in one system and something else entirely in another.

Unicode's entire purpose was to end that patchwork: replace hundreds of incompatible numbering systems with a single, universal one big enough for every writing system on Earth — used at once, not swapped in and out.

Think of Unicode as a single, agreed-upon reference book: every character used in writing is assigned exactly one entry, called a code point, written like U+0041. As of Unicode 17.0, that reference book holds 159,866 entries, and it grows every year as the Unicode Consortium — the nonprofit that maintains the standard — adds newly documented scripts, historical writing systems, technical symbols, and emoji.

It covers far more than letters: numerals from multiple numbering systems, punctuation, mathematical and currency symbols, musical notation, dead scripts no longer in everyday use, and yes, every emoji you've ever sent. That's why a "pizza" emoji sent from an iPhone shows up as the same pizza on an Android phone — both devices are pointing at the same Unicode entry, even though each draws it with its own artwork.

A code point is just a number in the registry — it doesn't, by itself, tell a computer how to store or send that number as actual bits and bytes on disk or over a network. That's the job of an encoding. Unicode defines a few (UTF-8, UTF-16, UTF-32), and UTF-8 is by far the most common on the web today.

UTF-8's trick is that it uses a variable number of bytes per character: common characters used in English take just one byte, while other characters take more, depending on how large their code point number is.

This is why a document's file size can shift depending on which languages it mostly contains, and it's the root cause of most "encoding errors" you've seen — they happen when software misreads bytes using the wrong encoding, producing the same kind of mojibake that plagued the pre-Unicode era, just from a different cause.

Unicode defines characters — the abstract idea of "the letter A" or "the Devanagari letter KA." It does not define how those characters look. That's the job of a glyph: the actual shape a font draws for a character (or sometimes for a whole cluster of characters combined) when you type it.

No font supports all 159,866 Unicode characters — type designers choose which scripts and character sets to support, then draw a glyph for each code point they cover. Some scripts make this straightforward: one code point, one glyph. Others — like Devanagari, Arabic, or Gurmukhi — require several glyphs per code point, because the shape changes depending on what surrounds it or where it sits in a word. That gap between "the character Unicode defines" and "the glyphs a font actually needs to draw" is the central design problem behind complex-script typography.

Once the basics click, the rest of this site is built for looking things up rather than reading start to finish.