>So it's _data-tag or _data_tag? again, i believe in discussion it's been _data-tag (following a general preference for hyphens, broken of course by the leading underscore); i'm not sure how _data_tag got in there. >(Note that _data_tag has the advantage that >it is a valid MCP ident but aussi a valid MOO/C/C++/Java ident... _data-tag >is not) true but i'm not sure how it's an advantage. --erik