in reply to Matching alphabetic characters

In my limited experience of locales, it is a bit of a nightmare. Especially when you want to have a multilingual websites where different characters are considered to be letters in different languags.

Also, when you use locale, you can't use \w to untaint input, because locale files are considered to be external to your script.

Anything to stop you making the leap to UTF8?

There is generally good support available for it now, and it handles multilingual sorting, etc etc.

I've found it the easiest standard to apply when developing web apps.