The heat from old fashioned light bulbs creates an interesting energy preservation perspective. We can cool the bulb with a fan and an aluminum cooling element even heat tubes can be used. We can also make an armature that isolates the bulb so that it's heat stays inside. Logically we should assume that extra heat extraction results in an equal increase in energy consumption. I'm not buying any of that until I see some figures.
Radiators are not useless, they do just that "radiate".
Candles are also very interesting in this context. I've done some experiments with a concept that basically exists out of a tea light with a pile of bricks over it. - It takes a long time for the bricks to heat up but when they are warm and you walk in and out for a bit so the room is cold again - the candle heat in the room is magically restored from the stones.
But how much heat is there in that little flame exactly? Is it all going into the air around it? How does it flow? Is it 1 thick strip of hot air straight up then it heats a circle of concrete on the ceiling?
I should try 2 long flexible metal tubes that fit inside another and attach that to a jar. then find out how much convection one candle can generate.
I imagine one can build quite a chimney labyrinth out of metal plates. The trick is to get a better mass/air balance and better heat distribution going on.
The flame will also burn slower if there is a lack of oxygen. If ventilation air from the outside is first lead into the candle heater we can make it so that the hottest air has the most oxygen in it. Who knows, there may be a lot more useful heat to be drawn from that little flame then we think. You should try boil water on it some time. It takes about 45 minutes to boil a liter.
It seems to me a simple heat exchanger can utilise a 40 watt bulb to heat an entire boiler. If the boiler is isolated the heat really cant go any place else.