You present some great questions. I would suspect that the ideal hook placement for salmon would be at the joint, and that allows the hook itself to be part of the structural equation as well if desired. The upfacing hook point is not an issue in Puget Sound and Straits waters. I need it on shallow flats to avoid picking up shells and debris, often putting the fly on the sand as part of the presentation but this is not an issue in salmon country, so it would be best to keep the hook on the same side as the eyes to swim upright and avoid fouling. A twopart body might be the ticket, with the hook coming out of the thorax about the same place the vent would be in a real candlefish (anterior limit of the abdominal cavity) and the back segment flexibly yet firmly attached.
With the right materials one could simply tie them both similarly and when chained together they would appear to be an extended candlefish fly; when the back is fished alone it would appear to be a shorter lance.
Have to think about this one, and tie up some prototypes.