Perhaps I should create an requirements outline after reading the source: From the post David links to: "2) Could somebody read the source and document the quirks of the parser in painstaking detail, so that somebody else can do a clean room implementation?" Gr, Danny