This in effect brings the pole pieces "in contact with each other" since the soft iron will be magnetically induced. Another way of putting it is that we short circuit the poles of the magnets. This will then ensure that the magnetic domains in the magnets will stay in alignment thereby keeping the magnet strong.
What do you think - should N poles be shorted with N poles or with S poles?
I think the N poles should be shorted with S poles. Because if it weren't shorted, the N pole would produce a field to connect with the S pole at its other end. Instead of doing that, it will send a field directly to an S pole of another magnet hence the short circuit.
Are you saying if a magnet is left by itself, the poles will eventually lose their magnetic ability because the alignment in the material whic makes the material magnetic will change to make N and S poles less magentic so as to lower potential energy of the poles within the same magnet? But by keeping another magnet close by the N and S poles are 'happy' being where they are and keeps them at a high potential. No change in alignment will occur since the N and S poles in this case are in separate materials and both ends are kept at high potential. The irons are present so as to disallow the magnets to touch? If they touch then the magnetic potential will die over time and two magnets will eventually turn into one magnet?
Yes, N poles should be shorted with S poles. To explain it it is best to refer to the magnetic field generated by the magnet. As we know the field fans out at the N pole and enters again at the S pole (bar magnet). The strength of the field keeps the little magnetic domains in alignment and keeps the magnet strong. Since the field fans out at the pole we find that an inverse square law describes the strength of the pole with distance from it. So by bridging the gap we in effect cause more field lines to re-enter the magnet and keep it strong. If the gap is not bridged external fields (earth) or mechanical shocks might cause some of the domains in the magnet to go out of alignment thereby causing the magnet to loose its strength.
So you are suggesting, it is the ends of the magnet that is responsible for the strengh of the magnet and by short circuiting it, we keep the end's field strong => strong magnet. Without the strong fields from the N pole, the domains in the magnet will weaken because there isn't anything to force it in place.
But one cause of concern is my looking at the magnet this way, we discount the sides. What happens if it's not rectangular but square? Than you would need 4 around it. That would be ideal even for a rectangular magnet woulnd't it?
With magnets we have a definite sense of direction of the field. The source of magnetism is twofold. Electron spin (up and down spin) and orbital motion of the electron. We find that with some atoms the orbital configuration and electron occupation is such that the magnetic contribution adds up thereby giving as an atom with a strong resultant magnetic field. In a piece of iron little regions (magnetic) form where nearby electrons orientate themselves such that we get a resultant magnetic field from that particular region. If more regions are coaxed into aligning in the same direction the field intensifies. The point where the field lines leave the magnet is called its pole and it is at this point that the magnetic influence is at its strongest (most magnetic field lines exist passes through this region). The field lines loop around and enter the magnet on the opposite side and at this point we find the other pole of the magnet. The magnetic influence is at their stongest at these two regions of the magnet. As we know a magnet can be designed such that these regions need not be at opposite ends of the magnet. This means that direction of the magnetic domains are bent along a curve from the one pole towards the other.
So if we are given a magnet with the shape of a square, we will have to do experiments to find where the two poles?
Is there a tendency for a piece of magnet to have electrons oriented such that two poles are created? Why does it tend to be along the far axis (or ends with maximum distance between each other) in a rectangular magnet?
Magnets are designed in all shapes and sizes. There is no reason why the poles cannot be on the short sides of a rectangle. The poles can be on any side of a cube, they do not need to be on opposite sides even.
Magnets always have two poles, N and S. This might have something to do with the fact that our motion relative to electric charge can be both in an approaching or receding direction - I do no know, but magnetic fields are apparently the result of our motion relative to electric charge (or vice versa).