Count characters

Four numbers, because "character" has four meanings

Take the emoji of a family, or a thumbs-up with a skin tone. A person sees one character. It is built from several code points joined by invisible connectors, so a regular expression sees seven. JavaScript stores text in 16-bit units and anything outside the basic range takes two of them, so it reports eleven. Encoded as UTF-8 for a database column it occupies twenty-five bytes. None of these numbers is wrong and none is the answer — the right one depends entirely on what is doing the limiting.

So the practical question is who imposed the limit. A database column declared as VARCHAR counts bytes in most engines, which is why a field that accepts 255 English characters rejects far fewer Japanese ones. An SMS is 160 characters in its own seven-bit alphabet, and drops to 70 per message the moment one character falls outside it — a single curly quote or emoji can turn one message into three. A form validator in JavaScript almost always counts UTF-16 units, which is why a field that says 280 characters may reject a shorter-looking string full of emoji.

The count of what a reader sees is the hardest one and is done with the browser's own text segmentation — the same machinery that decides where pressing the left arrow key takes your cursor. That is the right definition of "one character" for anything a person perceives, and it handles the cases that trip up simpler counting: an accented letter written as a base letter plus a combining mark is one, and Hindi and Thai clusters are counted the way a reader of those scripts would.

For ordinary English prose all four numbers are identical, which is why the problem stays invisible until it does not. They diverge on emoji, on accented and combining characters, on non-Latin scripts and on mathematical symbols — and those are exactly the inputs that cause a field to be silently truncated in a database three months after it was built.

When you want something else

In code, use a function that counts what you need rather than the default length. Intl.Segmenter in JavaScript counts what a reader sees; Python 3 strings count code points and len(s.encode("utf-8")) counts bytes; PHP has mb_strlen against strlen for exactly this distinction. The default in most languages is whichever the language found convenient, which is rarely the one your limit means.

When a database field is the constraint, the durable fix is upstream. Declaring a column as utf8mb4 in MySQL, or text in PostgreSQL, removes a whole class of truncation bug — the older utf8 in MySQL is famously only three bytes per character and cannot store an emoji at all. Counting carefully at the form is a workaround for a column that should have been declared differently.