cheap harmless (or nearly) lasers are everywhere writers are moving to other buzzwords. But I had no idea that anyone had come up with a more-or-less workable design until I stumbled across the "Stavatti SF-1" and "TIS-1".
(More below the jump.)
There are several good reasons why lasers have generally not been used in weapons in spite of their potential accuracy when shooting at moving targets. For one thing, standard lasers - for example the helium-neon designs that used to be standard in supermarket checkouts, or the diode lasers in pointers and optical drives - are fairly inefficient (only a small fraction of the energy in comes out as laser light). More seriously, standard lasers run on electricity, and it's difficult to come up with an energy storage device that can store much energy and deliver it in the brief, very intense pulses of electricity that would be required. This design avoids the problem by not running off electricity.
To make a laser lase, you need to get an atomic or molecular system out of thermodynamic equilibrium, so that photons can be amplified by the process of knocking atoms or molecules back towards equilibrium. This has traditionally been done by "pumping" the laser medium with light or electricity, but there are other approaches. One idea is that if you let a gas expand very rapidly, it cools rapidly. But while it's easy for the molecules' linear motion to slow down in such a setting, it takes time for the molecular rotation to slow down to match it. So you can get the molecules' rotational states out of equilibrium with their translational states. If you choose the right system, you can set up a laser based on returning these molecular states to equilibrium. This idea is called a "gas-dynamic laser", and it is interesting for high-power applications because it's relatively easy to set up very high-power gas flows. The most spectacular examples are rocket engines, which are in fact close cousins to gas-dynamic lasers.
My thought, for building a laser rifle on such a basis, would be to have it run on little chemical capsules containing fuel and oxidizer, which you combine to produce a blast of gas flow in which lasing occurs; you then eject the spent capsule and load another, very like bullets. But, perhaps because there simply wasn't enough energy available from such a system, the people designing this laser rifle went with a different idea.
Their system has a canister of mixed gases under pressure. This canister is kept hot by a rod of polonium, which is a strong alpha emitter (hence easy to shield the radiation from, though very poisonous if it ever got out). The hot gas is then used to drive the laser mechanism, expanding through a nozzle and flowing down a lasing chamber. The gas is then pumped back into the hot chamber by a pump powered by the radioactive polonium. So this rifle doesn't need any ammunition! Well, except for the polonium, which you have to replace every 90 days, whether you've been firing the rifle or not.
Will it work? Possibly. The gas-dynamic laser idea seems a bit weird, but such lasers have been considered for industrial use. Using polonium as a power source is unusual, but radioisotope thermoelectric generators have been used in everything from space probes to pacemakers. So I'm willing to believe you can make a portable laser rifle.
The design does have some problems, though. For one thing, the polonium is expensive. Nobody really knows how expensive, since it has only been produced in small quantities for research purposes, and each weapon would require kilogram quantities. Stavatti claims an initial cost of $2.6 million per weapon, though they also claim that this would drop to $15000 by 2018. There are also logistical problems; the polonium has a half-life of only some few hundred days, so Stavatti figures you'd need to replace the fuel canister every 90 days. Since the fuel canister is what costs the $2.6 million, that's $2.6 million (or if their estimates are good, only $15000 by 2018) per weapon every 90 days whether you fire it or not.
It seems to me, though, that the most serious problem is heat. The gas inside the fuel cylinder is at about 2000 K (and 300 atmospheres), so building the fuel canister so that this is something soldiers can sling on their bodies seems like something of a challenge. Even more serious, though, is the fact that the polonium is pouring out about 100 kW of heat at all times (i.e. whether you are firing the weapon or not). How on earth are you going to dump this much heat from a rifle-sized object? Surely it's going to have to have some incredibly hot radiator element, which will sear the flesh and set fire to almost anything that touches it. Worse, how do you store such a thing? A typical home furnace running flat out produces around 30 kW; simply storing one of these rifles in your house pours out three times that (which also says something about the energy costs of making one). Clearly you need to store them outdoors, and well separated. And, being (presumably) effective military weapons, under guard.
So, in short, I believe the laser mechanism can work, but to make it feasible they had to turn to a nuclear power source, which will be expensive and produce outrageous amounts of heat, and which you can't shut off. Energy storage is the key limitation here, as it is in many situations; and I think that underlying that is still the issue of efficiency: clearly you can store enough energy to kill someone in a small package, since that's exactly what a bullet does. But if you use a sufficiently inefficient way to transfer that energy to the target, your energy storage issues can become insurmountable.