I suggest you look at the thread titled how to associate the non xml data to xml data?
The problem is that you can not store any other xml data with non-xml data. The only way around this is to create two seperate xml documents, one xml and the other non-xml (they can both be in the same collection). You then store the document name or even url in the xml doc together with the other info/elements you need. You can then use this url to get to the pdf document or picture.
I agree, it is a bit cumbersome. Apparently the next version of Tamino will have an easier way of dealing with this.
I think the best thing to do for now is to create a servlet or dll that deals with your non-xml content as well as it’s “meta” data.
I am create unique prefix for non-XML docname (date, time, user, …) and
store full docname of non-XML part in URL in XML part.