What is the canonical way for a program to find out which character encoding to use for standard input, standard output, command line parameters, environment variables et cetera? One could look at the LANG environment variable, but if no encoding is given there? Is ISO 8859-1 the default for all languages? Björn Persson