I usually get feedback that I've mistaken some aspect of the Job Definition Format but attempting to describe it is one way to check how far understanding has gone.
The basis is a spec rather than a file format, although the spec is suitable for XML. PJTF- the portable job ticket format - is shown on the diagram but as a contribution to the current spec in the area between creation and pre-press. There is no assumption that the XML data will be stored as part of a PDF, so if a PDf is used it is just for the page description. The JDF will include all instructions. There is no intention to use Postscript code within PDF for device control.
There is reasonable confidence in XML being a standard. Triplearc offer one of the first products based on JDF and this is written in .NET. There will be a meeting for developers during GraphExpo in Chicago later this year. Members of CIP4 can uise C++ and Java code. This will be updated soon after the 1.1 upgrade later in April.
Creo fully support JDF and see 'Closed Loop - Open Systems' as a linked idea to 'Networked Graphic Production'. Agfa and Man Roland seem to be working towards an even better demonstration of JDF workflow at the next Drupa. Heidelberg made it very clear in the press conference that JDF will be fully supported in all their future products.
It would have been helpful for Adobe to have been there, just to confirm explanations on how this all fits with PDF.