• New functions matches, replace, and tokenise have a regular expression as a parameter
• Includes all the Perl regular expression features in XML Schemas plus some additions needed for tokenisation and replacement

Branches

• Regular expression consists of branches separated by |
• Each branch can consist of a set of pieces

Pieces

• A piece is an atom optionally followed by a quantifier
• Quantifiers are ?, *, +, {quantity},

Atom

A normal character matches itself and possibly its other case if a flag is set.

A character class such as escaped characters (\n etc), a character group such as [a-z], [^abc] (not abc) or a regexp

[0-9] [a-z]+

One branch consisting of two pieces matching a string consisting of a digit followed by one or more lower-case letters. + is the quantifier