Match, split, lex...

Multiple ways to work with regular expressions W3C XML Schema only uses regular expressions for a binary yes/no validation with matching. Developers who have used regular expressions in other contexts know that there's plenty more that can be done.
Match patterns Match patterns in scripting languages often return substrings from the match. While W3C XML Schema doesn't use this functionality, it supports the syntax (parentheses) needed to make it work.
Split patterns Many compound types use delimiters - white space, dashes, semi-colons, etc. - to separate the parts of the information they support. Regular fragmentations supports split approaches as well as match approaches.
lex-style work (not yet implemented) More sophisticated parsing can apply a series of regular expressions to content. This could prove very useful for difficult cases, but I haven't yet found a usable open-source processor which implements it in Java. Using lex-style processing might also require modification of the rules language.