XML_Transformer - character data hook

I'm looking at openoffice transforming (into good ole html) at present, and open office adds quite a few 'wierd' characters that need tidying up. and for some reason I've been getting some raw <? in some code examples. Anyway attached is a modification to add a characterData callback to the default namespace (eg. '' handler.) The other one that may be worth considering is the str_replace '-' with '_' for element methods. so <line-break> maps to start_line_break() I've got a few other ideas's that may be worth bouncing around on this. - changing the extends check to do a method check on namespace objects. - implement a base namespace that defaults to stripping tags rather than dumping them. - default to stdout debug logging rather than logfile - implement pear::error emits for non supported namespaces/tags.. - or some method to tally up and list this.. Otherwise it's great - A full open office to html convertor can be thrown together in hours, rather than the weeks it would take to understand XSLT :) Regards Alan