The controversy stems in part from the scope of the JDF spec—more than 500 pages. At that length, no one can quickly implement the whole spec. I previously criticized the complexity of the spec (see "In Search of PrintTalk and JDF," Printing Impressions, June 2001).
But I'm pulling back from that now. The spec is not written for the general public; it's written for developers. And no one is really intended to implement the whole spec—it covers too many processes for any single vendor (with the possible exception of Heidelberg). However it's the nature of JDF's design that vendors should create software systems that can process any JDF-formatted data, and extract only what's relevant to their own devices, rather than focusing solely on a subset of JDF. But it's also clear that developing such a software capacity will extend well beyond May of next year, so perhaps NGP is a practical alternative.