The original mXDE chose not to group with Document Consumer, but rather chose to add the mXDE actor to the ITI-43 (Retrieve Document Set) transaction. I recommend we change to grouping with Document Consumer (possibly Content Consumer).
I think we chose to do it this way so that the mXDE "Data Element Extractor" would not be expected to perform all the features expected of a Document Consumer. I think this is the typical Connectathon testing fear. A fear that the Connectathon monitors would force you to show user-interface features generally expected of Document Consumer. I think this is ridiculous gymnastics that we go thru in mXDE (and elsewhere) driven by our failure to define well how a Document Consumer should be tested. There are plenty of use-cases where Document Consumer is used purely by automation, if this is not testable then we need better testing.
The drawback of the current model is that it is unclear how this works with MHDS and XCA, as the editing is done to the XDS ITI-43, and not the MHDS/MHD ITI-68, or XCA ITI-39. Had we just grouped with Document Consumer, this support by MHD and XCA would be automatic without any additional changes to mXDE.
Further drawback is that the current model would seem to forbid use of DSUB/DSUBm... where if this was documented as a grouping with Document Consumer, this would be more open.