Hi, I'm interested to know how other applications handle string values input via browsers that do not send the encoding information. Without knowing the encoding of the input text, there seems no way that the correct Unicode string can be derived that reflects the text that the user intended. Our application uses servlet 2.2, and so is able to call ServletRequest.setCharacterEncoding(). But I don't see how this would help us anyhow, unless there is a way of determining the encoding of the input stream. The charset value in the content-type header of the *.jsp does appear to control the encoding of text input via IE. But Netscape appears to send the text in whatever encoding is specified as the "Character Set" value, regardless of the content-type header of the page. The user could be required to specify the langauge they are using, but I was hoping for something more dyanamic and reliable than this. This is probably not a struts specific issue, but any insights would be appreciated, Paul Darling