Abstract: A document is printed with graphical symbols which encode information but are not human readable; a graphical symbol has been selected from a set of visually distinct graphical symbols, which each code for the same bit or character, and a visible element printed on the document includes multiple such graphical symbols. Prior art graphical symbols (e.g. glyphs etc.) use a limited character set—e.g. just a single type of glyph for a type of bit or character (e.g. forward slash for bit ‘0’ and back slash for bit ‘1’). Hence, it is not possible to automatically select a particular glyph for use in one region of a visible element (e.g. a picture of a face) because it has an internal pixel arrangement that makes it best suited for that particular region for aesthetic reasons.