One of the advantages of DITA that was loudly proclaimed is that specialised topic files from one creator can be sent to another user and inserted into a document /without the specialised DTD being made available to that user/. The DITA OT will then produce some sort of readable (if badly formatted) output based on the parent element identified in the @class attribute. But it appears that none of the DITA-out-of-the-box editors which I have support specialisation in the absence of the specialisation DTD but barf in various interesting ways. Serna alone appears to make some attempt to support specialised elements. I have made a tiny tweak to <fig> to specialise it with a <credit> element based on <p>, and an <anchor> element based on <p>. This are subclassed in the approved manner, e.g. class=" topic/p topic/credit". It seems to me that a DITA editor on opening an unknown specialisation should be able to look at the inheritance tree in @class and temporarily rename the element to the nearest known ancestor. And, on save, it should write out the XML with the original element name restored from the leaf value in @class. As DITA is not a true object model and subclasses are merely renaming of ancestor elements for better semantic cueing, this does not appear to be a very difficult problem to solve. Anything would be better than the endless bleatings of Xalan error messages! I am basing my remarks on experience with Syntext Serna, <oXygen/>, XMLmind XML Editor (XXE), and FrameMaker v 8 with DITA-FMx 1.0. Perhaps XMetaL, Arbortext Author, and other editors already support this kind of compatibility. Regards, Hedley -- Hedley Finger 28 Regent Street Camberwell VIC 3124 Australia Tel. +61 3 9809 1229 Fax. (call phone first) Mob. (cell) +61 412 461 558 Email. "Hedley Finger" <hfinger@handholding.com.au>