Quite OMS, and I don't know where these ideas described by Chris came from but they are structurally odd to say the least. Assuming a beam is substantially evenly loaded the neutral axis (the bit which feels no forces at all) is along the geometric centre. Wooden floor joists are usually chosen for stiffness (in terms of the deflection they suffer when loads (people) move about on the floor above, and how this feels), rather than absolute bending strength to failure. With wooden joists the failure is a long way from significant deflection, unless the beam has a large knot near either edge, leading to large local deflection at that point. You have probably noticed this when trying to break up bits of scrap wood for a fire, even a bit of 2x1 (50x25 to metrics) is difficult to break across the 2" axis unless pretty long, but damage to the edges makes it much easier. Chris's diagram shows clearance around the ends where bearing takes place, which are subject to crushing but are very capable of supporting large loads as users of jacks and cars will have noticed. To see how big holes can be, examine the next steel building you are in. Often the roof beams are made by cutting a less deep beam along the centre in a strange pattern leaving round or hexagonal holes when welded back together with the correct offset. Thus one makes a lighter beam with less deflection and greater stiffness for given roof load. The same is true of trussed beams which are largely empty space across the centre, the load being on the two flanges, the upper one of which must be stiff enough between trusses to resist the compressive force. The bottom flange must resist the tensile load which is usually less of a problem. The notches are less satisfactory as they reduce the cross section exactly where the load is a maximum, and the stress is related to the square of the distance from the neutral axis, although cases of plumbers causing failure is very rare.  Holes in the centre of reasonable size (25mm) make an almost immeasurable difference to the strength of normal joists, simply because there is no stress there, but don't drill 100mm holes in a 120 mm deep beam and expect no effect! It is worth looking a a few older structures where you see the flange thickness is deliberately increased at the centre of spans, primaraly to reduce the deflection, rather than increase the failure load.

Quite OMS, and I don't know where these ideas described by Chris came from but they are structurally odd to say the least. Assuming a beam is substantially evenly loaded the neutral axis (the bit which feels no forces at all) is along the geometric centre. Wooden floor joists are usually chosen for stiffness (in terms of the deflection they suffer when loads (people) move about on the floor above, and how this feels), rather than absolute bending strength to failure. With wooden joists the failure is a long way from significant deflection, unless the beam has a large knot near either edge, leading to large local deflection at that point. You have probably noticed this when trying to break up bits of scrap wood for a fire, even a bit of 2x1 (50x25 to metrics) is difficult to break across the 2" axis unless pretty long, but damage to the edges makes it much easier. Chris's diagram shows clearance around the ends where bearing takes place, which are subject to crushing but are very capable of supporting large loads as users of jacks and cars will have noticed. To see how big holes can be, examine the next steel building you are in. Often the roof beams are made by cutting a less deep beam along the centre in a strange pattern leaving round or hexagonal holes when welded back together with the correct offset. Thus one makes a lighter beam with less deflection and greater stiffness for given roof load. The same is true of trussed beams which are largely empty space across the centre, the load being on the two flanges, the upper one of which must be stiff enough between trusses to resist the compressive force. The bottom flange must resist the tensile load which is usually less of a problem. The notches are less satisfactory as they reduce the cross section exactly where the load is a maximum, and the stress is related to the square of the distance from the neutral axis, although cases of plumbers causing failure is very rare.  Holes in the centre of reasonable size (25mm) make an almost immeasurable difference to the strength of normal joists, simply because there is no stress there, but don't drill 100mm holes in a 120 mm deep beam and expect no effect! It is worth looking a a few older structures where you see the flange thickness is deliberately increased at the centre of spans, primaraly to reduce the deflection, rather than increase the failure load.