Quite interesting stuff, we (Swinburne University) do quite a ot of experiments with it in the hope of understanding what new physics you can achieve. One example is that we hope to b able to create a atomic interferometer which requires a large number of atoms to behave as a single wave (a normal interferometer uses the coherence of laser light as all the photons have extremely similar properties which bahave like a single wave). In a BEC, all the atoms have the same quantum state, so if you know your quantum physics that means their wavefunction is the same, making one single wave. The bennefits are that the wavelength of the atom wave is below any light that we can produce coherently with any ease. This means more accurate measurments, far better than the extremely accurate measurments made currently with optical interferometers.
Just recently, we managed to go one step further and generate a mollecular bose einstein condensate, pretty much the same thing but with molecules rather than atoms.
http://swinburne.edu.au/feis/caous/news mbec.htm
As far as lasers go, they are fairly stock standard scientific lasers (continuous wave I believe) that lase at a specific wavelength that the atoms absorb. The one key when using them for cooling is to be able to tune them incredibly precisely, and to minimise any drift, such as that caused by vibrations in the bilding the laser is housed in.
The trick to using them to cool down atoms rather than heat them up is to use their small amount of momentum to act against the atoms movement as mentioned by Tallaxo. To do this, you tune the laser to sightly above the absorbtion frequency of the atom. Any atom moving towards the laser beam will see the normal laser frequency, but will add a small amount due to its movement towards the laser. The same reason why a car horn changes sound from high to low when passing you on the highway. The atom which would normally see an "invisible" photon in the laser beam would now see a "visible" photon that it will absorb. The conservation of momentum means that it will slow the atom down so that it no longer sees a "visible" photon. If you put that laser through your atom chamber in three dimensions, and two directions (back and forth) for each dimension then this process works to slow down atoms moving in every single direction. You start with the laser tuned an appropriate distance from the atoms absorbtion to cool the fast atoms, then slowly tune it closer to the atoms absorbtion frequency so that you continue to cool down the atoms to about 1/10000th of a degree above absolute zero.
You are then faced with the issue that even cold atoms move out of the intersection of the laser beams and heat up when they come into contact with the sides of the chamber or other hot atoms. To solve this, you use vacuum pumps to remove all unwanted atoms and create a specific magnetic field that changes the properties of the atoms in such a way that atoms in the middle of the laser trap are not effected, but any atom that tries to move away experiences an increased force pushing them back into the middle.
You now have pretty damn cold atoms in what is called a Magneto Optical Trap (MOT). The atoms are still far too hot though to create a BEC though, and to cool them further you need to remove the MOT and increase the magnetic field so that only the magnetic field is holding them in place. Now, all the hot atoms will slowly boil out of the magnetic trap with the effect of cooling the remainder. Using this technique you can cool down the atoms to the temperature required for a BEC.
For a great explanation on the topic along with some interactive applets, check out
http://colorado.edu/physics/2000/bec/lascool1.html
As for the slow light, I have worked on tis very topic myself, only with warm Rubidium. One important fact that needs to be pointed out is that techniques like these do not slow down the phase velocity of light, that still remains at 3x10^8 for free space and slower in materials with hgiher refractive indexes. What these techniques do, it use atomic properties to modify the optical property of dispersion so that the group velocity of light slows down, speeds up or goes in reverse. What this effects, is the "envelope" of light, a pulse of light on a laser beam for example. So, the light still travels at full speed, but it behaves in such a way that the pulse of light travels slower, faster or ackwards.
Slower has many uses in optical buffers. Currently, all information processing is done electronicly with the majority of transmission done optically. To improve performance of the information processing, it is best to try and perform the processing optically, not electricaly. To do this however, we need to be able to store photons as we do electrons. At the moment, this is quite hard but this technique of slowing down a pulse of light can help us do it. As for maing a pulse go faster, this is more difficult as Einstein stated and is yet to be proven wrong, that information can not travel faster than the speed of light. Experiments with "Fast lgiht" so far have shown this, as the pulse is significantly distorted as a result of traveling faster.
http://galileo.phys.virginia.edu/classes/109N/more_stuff/Applets/sines/GroupVelocity.html is some informaton on the group velocity of light. Set the frequency to high and take a look at the speed the small peaks move to the right. This is the phase velocity. Think of this as the waves of your laser moving from the pointer to the wall. The group velocity is the large (wide) peaks. You can change their speed so they travel slower or faster than the speed of light. In terms of your laser, you would see this as a pulse (not on or off) but say an increase in intensity moving from the pointer to the wall. This is what we can change to be slower, faster of backwards, but currently not in air so dont be expecting that you will be able to outrun a pulse of light any time soon.
Hope that is informative for all of you.
****EDIT Fixed Links****
that's what hit me the most in what you said :O don't know enough about light to know what makes it accelerate again when it comes out 
