Personally for me, screen real estate of my controls for VDMX is at a real premium.
In the case of my QC compositions that generate their own video without tapping any video in VDMX, I have found that with a lot of controls I need to open up a very large Layer Controls window. When I go back to using a movie or image in that layer, I have to then resize back to the default small size. Also while this window is expanded it covers up my efx for that layer. With that all said, I decided to see about running the composition as an efx rather than a comp.
I found that if I added an imageInput splitter and hooked it up to a disabled billboard, VDMX would recognize it in the efx list and load it. You only get the output of the QC and not any loaded media on the layer. You do have to load something for this to work- I just add a still image to that layer to use minimum resources.
The advantage now is I can have this QC image generator running out of the efx chain portion of the layer, which for me is much tidier to administer. I realize this is sort of going against the design philosophy, but other than that I'd like to see if the developers could comment if they see any major performance or other related issues with running a QC this way rather than loading it as a composition.