To me this is really a nusance: some Chinese web pages are in UTF-8 (unicode), some are in GB2312. Internet Explorer can not determine the Chinese code format from a web page header and makes a wrong guess. Basically the user has to go to View => Encoding and try those two formats to see if it will work.
This is not an isolated problem for Chinese language, or for web browsers. Similar problems do occur to other non-western languages, and software. Before we have a better solution in the software, I wonder if we could all agree on UTF-8, which seems to me is a newer (and better) standard.