Indeed, force is applied on surface, as well as weight is applied to every particle. But in sum, force is the same as if it would be applied to single point.
As about disregarding pressure difference, damnit, as all buoyant force is result of pressure difference, don't you know that?

In hydrostatical case (that is, no turbulence/something), force is applied to center of mass of displaced volume of water, and if i understood arithma's post right, exactly as guessed.

It is very easy to prove ala Archimedes theorem. Water in water doesn't spin, and there's no turning force. All forces is compensated. If we have water in the bag, we can assume in computations that weight is applied to center of mass. And bag with water doesn't turn in water, therefore buoyant force neutralizes weight, and to do that it have to in sum be applied to center.

In summary:
1: Sum of forces applied acts exactly as one force applied to center of mass of displaced volume of water. (well, almost exactly, if you gonna to take curvature of Earth and non-linearity of gravitation field into account, you'll get some extremely small difference)
2: Motorherp's idea about disregarding pressure difference is completely incorrect.

Quote:Original post by Motorherp
My bad. I had it in my head you were refering to applying the force at the objects centre of mass. Totally missed the submerged volume centre of mass thing. However, computationaly speaking, unless you are dealing with very simple shapes you'll find it difficult to approximate this 'submerged volume' and harder still to find its centre of mass. Hence for complex shapes you might be better sticking with an idea like I said where the buyancy force is approximated at points distributed accross the submerged surface.

Quote:Original post by Dmytry
2: Motorherp's idea about disregarding pressure difference is completely incorrect.

Not exactly. What I was refering to is that in most cases the pressure difference between two points on the bottom of the submerged surface, where in my example the force is approximated, is negligable.

not neligable. Buoyancy force is result of pressure difference.

And if you properly compute surface integral of force (for every submerged triangle of mesh), it will be the same as computing submerged volume and center of mass of that volume. It is relatively easy.

if you gonna to make any physics, you anyway will have to do many things on per-triangle basis, and will have to have low-poly models for physics anyway. So probably not that costly. Integral for every triangle is analitical, you need to check every triangle of mesh, and pseudocode:

Vector3 sum(0,0,0);
Float distlaced_mass;
for every triangle{
if triangle is completely underwater, compute effect of volume of water enclosed by this triangle and it's projection onto water surface, kinda prism-like. That is, find mass of water in that volume and add to displaced_mass, find center of mass of that volume, add coordinate of center of mass multiplied by mass to the sum
else
if triangle is partially underwater
cut triangle with water plane leaving unly underwater part, do the same as above with that cut surface.
}
center_of_mass_of_underwater_volume=sum*(1/total_mass)

Note: For clockwise or counterclockwise looking-from-top triangles, volume in calculations must be negative. That way, it should work even with concave mesh.