Re: Latest proposal for smart regexes in RELAX NG

> > No. This is so always. Because the locale only matters when documents
> > are read or written. When documents are processed (validated) or
> > matched against Unicode regular expressions, the encoding is always
> > Unicode.
> 
> In XML Schema regexes, yes; not in all tools using regexes today.

We are discussing application of regular expressions to Relax NG
validation, aren't we?

> 
> > That's what I am saying; exactly for the same matter there is no
> > string-regex-ish formulation for XML, there shouldn't be xml-regex-ish
> > formulation for strings.
> 
> Regular expressions aren't strings: they have nesting.

Strings may have structure.  A good example is text in human language.