My understanding is that the ALB cross-section (rectangle) is more efficient from an engineering standpoint, assuming that the bending beam is of uniform material. So for a selfbow made entirely out of, say, blacklocust heartwood, a rectangular cross section will place the limbs under less stress for a given force applied compared to a D-section longbow. You would expect the d-section bow to take more set because more compression force is being exerted on the belly of the bow due to its depth. The set robs early draw-weight and, consequently, cast.
However, if the material is not uniform, as in a yew sapwood-heartwood ELB or a laminated bow like a BBI, a narrow, deep section can work better. The belly material on these bows can handle more compression force than the back material, and vice versa. An additional advantage to a narrow deep design is that it weighs less for a given stiffness (draw-weight) because bend resistance increases by a greater factor per additional thickness than for additional width (i.e. doubling the thickness of a beam increases its stiffness more than doubling its width).
So, the answer is "it depends". A lot of people seem to have good success with ELB-type cross sections with whitewood selfbows, which is sort of counterintuitive to what I just said above. I suspect the performance comes from the narrowness and lightness of a deep-section bow, and possibly better dynamics.
There are other aspects, too, that should be considered in a comparison, like ease of construction, stability, archer's paradox, etc. At any rate, to broadly say that ALBs are better than ELBs is ignorant. You need to more narrowly define the comparison in terms of goals, materials, etc.
-Rob
