Julian Reschke wrote: > I'm not sure about the futile part; AFAIU, there are valid use cases for > the above. Sure, IBM's format seems to work for IBM's use case. I just think that trying to define a generic, application-independent mapping between JSON and XML is not so useful. > But XML->JSON of course is much more complex; if a generic > transformation is needed the result will be extremely ugly. Obviously. The XML data model (e.g., ESIS) is much more complex than JSON's, and whatever generic mapping from ESIS to the JSON data model is defined, it won't "look like JSON" in the end. Grüße, Carsten