When I'm measuring my food, I like to shake it, so I can fit more in the measuring cup without increasing the calorie count. ;) Oatmeal will pack down quite a bit with some shaking.
The title is funny. I realize it's sometimes hard to translate and summarize scientific results for the layperson, but the paper wasn't comparing stirring or shaking. They compared rotating the container to tapping the container. Rotating is, ironically, a type of shaking, and closer to shaking than tapping. I would expect a good deal of dice packing in this experiment using vertical shaking, as opposed to lateral tapping.
Vertical shaking would unpack just as many things as it would pack. Horizontal shaking works because gravity is biasing the reorganization toward more packed arrangements.
Maybe you're thinking of (the equivalent in a crystalline solid) of "boiling"? E.g., giving a bump on the bottom of the cylinder, that would have enough force to scatter individual cubes, but not enough force to unlock cubes from the crystal.
It sounds like you're imagining a difference in magnitude, as opposed to a difference in type of motion, you seem to be assuming that I'm suggesting a very large magnitude. I would call bumping the bottom a type of vertical shaking, if the container is allowed to move vertically, regardless of how far (or short). A vertical shake that moves less than the height of one die doesn't move far enough to allow unpacking, and empirically it works for me. Fast and small vertical shaking, or boiling, if you prefer, will "liquefy" the particles and allow them to settle. Fast and small horizontal shakes would probably do the same, I just speculated above that vertical boiling packs faster than horizontal "tapping", but who knows, I could be wrong, I'd have to do the experiment.
Hmm, good point; I wasn't considering <1.0h shakes. (I was mentally considering anything worth being called a "shake" as raising and lowering the entire packed column between the top and bottom of the cylinder, such that it would repeatedly bump both the top and bottom.) A 0.5h shake is indeed effectively equivalent to a Newton's-cradle-like bump.
The title is funny. I realize it's sometimes hard to translate and summarize scientific results for the layperson, but the paper wasn't comparing stirring or shaking. They compared rotating the container to tapping the container. Rotating is, ironically, a type of shaking, and closer to shaking than tapping. I would expect a good deal of dice packing in this experiment using vertical shaking, as opposed to lateral tapping.