Running a fuel tank dry
I've recently read some insightful articles on piston engines by Avweb columnist John Deakin. He advocates the practise of running a fuel tank dry in-flight (as long as another tank contains fuel) and vehemently dismisses the notion that doing so can damage the engine and/or components of the fuel system.
I'd really like to know the mechanical implications of doing this and the answer to the question of whether not it has a quantifiable negative effect on the engine. Letting combustion stop once you're airborne is, to me, rather bold - especially when you consider that the turning of a knob five seconds earlier would have kept the fire going. I am however of the opinion that behavioural principles should result from something tangible rather than an overly-cautious outlook.
If the consensus is that running a tank dry is harmful, how much fuel needs to remain for satisfactory function? Would it be any less harmful to change to a fuller tank just as a reduction in fuel pressure is indicated, or would the hypothetical damage have already been done at this stage?
Many thanks.