A client has localized an English ndx.xml file that contains a JavaHelp index into Korean. This is the JavaHelp index file. He has set the character encoding to UTF-8, and saved it from Notepad as type UTF-8. Using this file, the JavaHelp displays correctly, and the index displays correctly. I can enter an English term in the index navigator and it finds the entry and topic just fine. (Some terms are untranslated.) However, if I enter a Korean string, the help system hangs and CPU usage goes to 100%. I have to kill the JavaHelp with Task Manager. I feel that the problem is somehow related to how the file was saved in Notepad. I have a similar file in Chinese that works properly. It is saved as ANSI. The characters are not readable in Notepad, but the index does function properly with Chinese characters from the JavaHelp. I can provide the files (Korean, English, and Chinese) if it would help.