Currency symbols

A symbol is not a currency

This is the practical problem with currency signs, and no amount of copying the right character fixes it: several signs are shared by several currencies.

So $50 is not a price until the context says which dollar, and ¥1000 is about US$6.70 in Japan and about US$140 in China — a factor of twenty.

When to use the ISO code instead

ISO 4217 gives every currency a unique three-letter code, and the first two letters are the country:

Use the symbol where the currency is obvious from context — a price on a Taiwanese shop’s own site — and the code where it is not: invoices, contracts, anything crossing a border, and any interface a stranger might read.

The common disambiguating prefixes (HK$, NT$, US$, A$, S$) are conventions rather than standards, but they are widely understood and better than a bare $.

Where the symbol goes

Placement is a local convention, not a property of the currency:

The euro is the notable split: English-language usage puts it first (€50) and most of the eurozone’s own languages put it last (50 €). Both are correct in their own context.

Full-width and half-width

CJK text has full-width forms of the common signs, which sit correctly on the character grid:

Inside a line of Chinese or Japanese text the full-width form spaces correctly; in Latin text it looks oversized. Same rule as the rest of CJK punctuation.

Note that ￥ (U+FFE5) and ¥ (U+00A5) are different characters. A price list mixing them will not sort or search consistently.

The ones you may not recognise

The generic ¤ is the fallback a system prints when it has a number formatted as money and no currency to attach — seeing it in production usually means a locale was not configured.