swa66 - 6:25 am on Sep 26, 2012 (gmt 0)
AFAIK, there's a problem with IE (it's always the same culprit isn't it?) in that when it tries to encode things in iso-latin-1, it sometimes keeps characters in UTF-8. Trying to force it all to stay in iso-latin-1 is futile once your input braks up like that.
ve never seen that wacko behaviour if it's all moved to UTF-8 and make sure not to do conversions anymore (it's safe to convert between unicode encodings (UTF-8 to and from UTF-16 is ok). [you use UTF-16 for some databases (mysql does UTF-8 - but e.g. MSFT's SQL server uses UTF-16 internally]
IMHO: the sooner you move it all to UTF-8 the sooner your can stop messing up your data.