This is an archived snapshot of W3C's public bugzilla bug tracker, decommissioned in April 2019. Please see the home page for more details.
In the introduction of the "Tokenization" section, it is said that a number token has a representation of *1 or more* characters: > Number, percentage, and dimension tokens have a representation composed of 1 or more character... However, in the "Number state" section, the representation is set to an *empty* string: > Create a number token with its representation initially set to the empty string.
Yes, it's created with an initially empty representation, but it's never emitted with one. Those definitions talk about the tokens you'll actually find in the token stream, not temporary ones under construction by the tokenizer.