Our site is having severe networking problems right now, so I cannot check anything in to day. I have been moving many of my existing import tests into the xml test framework, and have basic (non-namespaced) importing working against them. As of this afternoon the hp implementaton can : -import .cdl files with absolute file:, http: URLs. There is no security here. -import .cdl files with a classpath:/package/package/filename.cdl URL. the classloader will only load signed JARs, so this is the secure way to load things. -do relative imports. If a file comes from http://example.org/a/b/example.cdl and it imports ../c/ex2.cdl, then the URL will be http://example.org/a/c/ex2.cdl. Same goes for classpath: imports. -do relative imports from imported documents, as with these two from a forthcoming test case <ct:document uri="http://cddlm.org/chain-target.cdl"> <cdl:cdl> <cdl:configuration> <WebServer> <hostname>localhost</hostname> <port>80</port> </WebServer> </cdl:configuration> </cdl:cdl> </ct:document> <ct:document uri="http://cddlm.org/chain-source.cdl"> <cdl:cdl> <cdl:import location="chain-target.cdl"/> </cdl:cdl> </ct:document> </ct:documents> -detect recursive imports and halt the process gracefully. SmartFrog classic doesnt do this incidentally; it runs out of stack space, which is an implicit failure, and not so good for diagnostics. I will be checking in tests for all of these, with the recursive ones marked as required=false as the spec says nothing about how well to handle recursive imports. We may want to mark everything with relative URLs as required=false too, as again, there is nothing explicit there.