Theory-wise, gauss guns and rail guns are perfectly symmetric, because the relationship between voltage and current is identical to the relationship between electrostatic and electromagnetic forces.
In either case, the purpose of the gun is to convert electrical energy into kinetic energy through some intermediate energy storage field, either electrostatic or electromagnetic.
The basic idea behind the electrostatic gun (rail gun) is that a charged particle, allowed to drift in an electric field will "fall" in the direction of the field, extracting energy from the field in the form of kinetic energy as it does so.
In the electromagnetic gun (gauss gun), the basic idea is that any particle which bears a current, when allowed to drift in a magnetic field will "fall" in the direction of the field, extracting energy from the field in the form of kinetic energy as it does so.
The major hurdle arises when this particle reaches the end of the field. At that point the field will likely reverse direction or eddy or something else that would make the particle immediately start slowing down.
In both the gauss and rail guns, the solution is simply to make the field "follow" the projectile, though the reason this works in both cases is very different.
In the case of the rail gun, you've got a charged particle falling from one plate of a capacitor to the other. You've got to have a hole in one (or both) plates so that when the particle reaches the far side of the capacitor it can pass through. But now the field reverses direction (because it's always pointed towards the most negatively charged plate), and that makes the particle slow down. On the other hand, if (just as the particle passes through the plate) we reverse that plate's charge, then we build a new field which is still in the correct direction to accelerate the particle.
In the case of the gauss gun, you don't want to make your projectile carry a current: that involves giving it a power source, and that drives your projectile cost through the roof. So instead you use the magnetic field of your gun to generate eddy currents in the projectile. The problem is that the force generated by the eddy current is dependant on whether the projectile is travelling through an area of increasing or decreasing field intensity. You use coils in the gauss gun to generate your magnetic field, and let the projectile fall through the centers of the coils. As the projectile approaches one coil, the field lines are getting squeezed closer together to fit through the center of the coil. That's an increasing field density (positive flux), so the eddy currents are going the right direction to accelerate the projectile. When the projectile is in the middle of the coil, the field is constant (0 flux) so the eddy current isn't generating any force on the projectile: it's just coasting. On the far side of the coil, the field density is decreasing as the field lines are no longer constricted by the geometry of the coil (negative flux) and the eddy current is now generating a force that slows down the projectile. So while the projectile is coasting through the coil, you switch the field direction. This allows the projectile to continue accelerating as it leaves the coil (because now the field, while still decreasing, is in the opposite direction. The hard part is in switching the field direction without making the projectile slow down. There are ways to do it, but it's difficult.
So in both the gauss gun and the rail gun, you start with the projectile at one end of the barrel with all your field generators (electrodes or coils) set to the same polarity (all electrodes positively charged, or all coils supporting a clockwise current). As the projectile passes each field generator, that generator switches direction. When the projectile finally leaves the barrel, all the field generators are set to the opposite polarity of when they started (electrodes negatively charged, all coils supporting a counterclockwise current).
It's very easy to detect the position of the projectile in the barrel, and synchronize the field reversals with that position: each of the field generators also acts as a sensor, feeding an appropriate signal back to the power supply as the projectile passes through. This signal can be used to trip the field reversals.
The reversals, however, are quite hard. In the case of the electrodes, you're trying to dump all the charge on one electrode quickly and add an equal (but opposite) charge back on. In the case of the coils, you're trying to stop a large current and add another in the opposite direction. In either case, you're expending an energy that's twice the energy stored in the field that the generator is supporting.
To quickly dump the charge on a capacitor and replace it with an opposite charge, you need to use a large current. To quickly stop the current in a coil and replace it with the opposite current, you need a large voltage. Are you beginning to see the symmetry between the two devices now? I sure hope so.
A year ago or so, I came up with a circuit that can perform this field reversal very simply. It's probably patentable. Nevertheless, if anyone asks me directly to my face, I'd probably be willing to explain the circuit. Hmmm. Maybe I should talk about it at DefCon? We could do a joint talk between Ming and I. *shrug* You reading this Ming?
So what all this means is that for the gauss gun, you need a power supply that can support a long term large current with high voltage transients. For the rail gun, you need a power supply that can support a long term large voltage with high current transients.
BTW: A rail gun gets its name because the science fiction story that originally proposed the gun had the electrodes mounted on twin rails between which the projectile travelled. A gauss gun gets its name because the power of the gun has a lot to do with its field strength, which is measured in Gauss.