You can perfectly well have all radiant heat extraction instead of the typical mix of radiant barrel and thermal storage mass. This is recommended practice for spaces like shops which are used intermittently and don't need to stay warm between uses.
Be aware that it is not good practice for a continuously warm space, as trying to damp down the burn rate to avoid overheating in other than the coldest weather will cause inefficient combustion and creosote buildup, just like an ordinary woodstove.
The big thing to remember is to have a highly insulated combustion zone so you get complete combustion and elimination of pollutants, followed by heat extraction. For a large space, you can use a stack of two barrels for a big radiator.
This
thread has a lot of discussion about massless
heaters, with significant inputs by Erica Wisner, one of the most experienced
RMH builders in the world:
https://permies.com/t/14609/wood-burning-stoves/RMH-flue-fires
Addressing most of your specific questions:
You can do all kinds of fancy experiments with reburners, but you
should build a standard
RMH by the book first so you have an idea what is achievable without fuss. You can't know if you have an improvement without knowing what you are trying to improve. Generally, a well-built J-tube with P-channel and tripwire (two simple enhancements you can google for details) gets close to as efficient as it is physically possible to be, and it would be a stretch to do better without considerable skill and experimentation.
A metal enclosure is fine, as long as the surfaces exposed to flame or high heat are refractory/ceramic. Apparently the guy in the video used to build all-metal units, and has taken to lining them with insulating material.
An underfloor mass/duct system is possible as long as the rest of the design follows proven principles and proportions. You do need a good chimney to ensure constant draft. 10' would probably be a minimum for effectiveness, and it needs to end at least three feet above the roof where it exits, and two feet above any roof within 10'. These are code minimums, and in some cases might need to be increased for good performance.
The smaller a RMH system diameter is, the more surface-to-volume and drag effects degrade its performance. 6" and 8" are common, easily-built sizes. 4" requires good siting and very careful construction, and still might not work well; it is considered an advanced build. Smaller than 4" has not been found to be viable for J-tube systems. 10" or larger is a serious beast, only appropriate for heating very large spaces or extreme climates.
A J-tube gets several times more air than needed for full combustion, so improvements will not come from introducing more air, but from better turbulent mixing. The simple P-channel and tripwire, added to the sharp bends in flame path in the J-tube, do this very well.
The position of the exhaust from the barrel base relative to the burn tunnel is immaterial, as long as you have good smooth airflow. It can go in any direction you wish.
Plenty of people have built
greenhouse RMHs with fish or other water tanks above the buried duct. You just need to ensure that the water cannot get above boiling without an easy, NON-PRESSURIZED, steam escape path in case of overheating.
Waste oil has been used as fuel, but requires special handling. Search to find a number of experiments with it.