I have found a bug or undocumented limitation in the boost serialization library. I'm using VC++ 2003, which allows you to use variable names which include dollar signs ($). Variables with names containing dollar signs can be saved to an archive using the boost::archive::xml_oarchive class. When boost::archive::xml_iarchive is used to reload the variable, however, the parsing of the variable name fails, and basic_xml_iarchive<Archive>::load_start(const char *name) calls boost::throw_exception( archive_exception(archive_exception::stream_error) ). For example, std::string filename( "C:\\dollar.txt" ); int $(0); { std::ofstream archive_file( filename.c_str() ); if( archive_file.good() ) { boost::archive::xml_oarchive output_archive(archive_file); output_archive << BOOST_SERIALIZATION_NVP( $ ); } } produces the following file: <?xml version="1.0" encoding="UTF-8" standalone="yes" ?> <!DOCTYPE boost_serialization> <$>0 but trying to load the file causes a boost::archive::archive_exception to be thrown: { std::ifstream archive_file( filename.c_str() ); if( archive_file.good() ) { boost::archive::xml_iarchive input_archive(archive_file); input_archive >> BOOST_SERIALIZATION_NVP( $ ); // throws } } I think the spirit definitions for the names of variables are in C:\boost_1_32_0\libs\serialization\src\basic_xml_grammar.ipp and C:\boost_1_32_0\libs\serialization\src\xml_grammar.cpp. Name = (Letter | '_' | ':') >> *(NameChar); and template<> void xml_grammar::init_chset(){ Char = chset_t("\x9\xA\xD\x20-\xFF"); Letter = chset_t("\x41-\x5A\x61-\x7A\xC0-\xD6\xD8-\xF6\xF8-\xFF"); Digit = chset_t("0-9"); Extender = chset_t('\xB7'); Sch = chset_t("\x20\x9\xD\xA"); NameChar = Letter | Digit | chset_p("._:-") | Extender ; } ______________________________________________________________________ This email has been scanned by the MessageLabs Email Security System. For more information please visit http://www.messagelabs.com/email ______________________________________________________________________