The Infinity Symbol

The character

U+221E INFINITY is general category Sm, a mathematical operator — the same class as the plus and minus signs, not a picture. Its shape is a lemniscate: one closed curve that crosses itself once. It is not a rotated digit eight and should not be made by rotating one, because a rotated 8 is still an 8 to a search index, to a copy and paste, and to anything that reads the text rather than looks at it.

It is three bytes in UTF-8, so a three-byte column stores it. In HTML it is ∞ or ∞. Its East Asian width is Ambiguous, and it is the only one of the five that is, so a terminal set to render ambiguous characters as double-width gives it two columns and the other four one.

The other four

U+29DC INCOMPLETE INFINITY is the curve drawn unclosed. U+29DD TIE OVER INFINITY carries a tie mark above the sign. U+29DE INFINITY NEGATED WITH VERTICAL BAR is struck through, which is how Unicode writes the negation of an operator. U+2BF9 EQUALS SIGN WITH INFINITY BELOW is the odd one out: its category is So rather than Sm, so a rule that selects mathematical operators picks up the first four and misses this one without saying so. All five are three bytes, and the four beyond U+221E are Narrow.

Characters that look like it and are not

U+221D PROPORTIONAL TO sits one codepoint below INFINITY and is an open loop that is easy to mistake for it at small sizes; it is also Sm and it means something else. U+267E PERMANENT PAPER SIGN is an infinity sign inside a circle, category So. There is also a combining infinity mark at U+1AB2, which is not a standalone character at all: it attaches to whatever precedes it, so pasting it alone decorates the character before it instead of appearing by itself. None of those three is in the grid above. When a document does not look the way you expect, the codepoint is the way to find out which character is actually in it.

Where it breaks

• A field that strips non-ASCII removes it, and there is no ASCII substitute that means the same thing — two letter o’s and a lowercase x are both something else.
• It is not a number to any code that parses numbers. A field expecting a numeric value rejects it, and a cell holding it holds text.
• A font without coverage in the mathematical operators shows a box; the character is still there and still three bytes.

Nearby

This one belongs to the maths wing: maths and science symbols lists every set in it. Written into an interval it sits beside an inequality sign, and those are on inequality symbols; the rest of the operators are on math symbols.