Errors
In the words of Homer Simpson, "Doh!"

Some late night errors in the algorithm I described yesterday (for a start, that one wasn't tail recursive). Out of embarressment I went back and changed it rather than providing an addendum. It's probably still wrong, but it's closer. :-)

For the benefit of AN: I have a lot more unit tests for JRDF. I've tried to stick to interfaces mostly, to make it adaptable to other JRDF implementations. I've had to use SOME JRDFmem specific code, but perhaps I can break that out into a subclass.