A mass launcher is any device that shoots a payload into space through a launch system based on the ground. In other words, a giant space gun that shoots shipping containers. A mass launcher that uses magnetic force to launch (like a railgun) may sometimes be called a mass driver. One that uses projectiles fired from the surface to accelerate the payload is usually called a space fountain.
The most common hypothetical forms of mass launcher are ones that uses magnetic propulsion or compressed gas to launch a capsule through a long tube with enough velocity to exit the atmosphere. The original1 mass launcher, put forth by Marshall Savage in his book The Millennial Project, was dug into the tallest of equatorial mountains, to take advantage of the thinner atmosphere at higher altitudes and the additional velocity provided by Earth's spin at the equator. It used magnetic acceleration and was expected to start out horizontally, curving towards the vertical as it approached launch speed.
There are a number of problems with this design. The shorter the accelerator track is, the higher the g forces will be; Savage's planned 250 kilometer track still would have created g forces strong enough to kill a human (up to 250 G at the sharpest part of the curve). If we can construct a sufficiently long track, we can reduce the g forces, but any curve will increase g-forces around that curve. Of course, this does not mean that a short, curved track would be useless; there are a lot of things that we would like to put up into space that can take severe stress. Tunneling through mountains is not cheap or easy, and each tunnel dug would obviously be pointed at exactly the same spot -- a lot of effort for a single launch trajectory. And the difficulty of getting the necessarily building permits has also been cited as a problem; there are only three mountains that meet the criteria (if we limit ourselves to the equator): Mt. Kilimanjaro, Mt. Cayambe (an active volcano), and Cotopaxi.
The First Millennial Foundation/Living Universe Foundation, finding that building a mass launcher on equatorial mountains was not practical, shifted their focus to marine mass launchers. Oceanic launchers have some distinct advantages. First off, the oceanic launchers would be straight tubes, without the drastic increase in g forces that results from the curve in the montane tube. And while hardly agile, a 'floating' launcher could be adjusted to point in various directions. The ocean also makes an acceptable landing pad for return payloads, making round trips more feasible. Of course, the difficulty of keeping the launcher rigid and immobile during launch would require as-of-yet-undiscovered technologies, and together with the difficulty of powering hundreds of kilometers of track, it starts to look like a space elevator might be easier, if not cheaper. However, a company called QuickLaunch Inc.2 has proposed a marine mass launcher using hydrogen gas.
There are other possibilities. As recently as 2007, DARPA was still experimenting with a device called a 'slingatron', a large construction that is based around a circular or spiral tube in which the payload would accelerate as it traveled the length of the tube. This coiled tube is much more compact than the earlier plans for tracks 100's of kilometers long. The slingatron has the added wrinkle that it does not accelerate the payload through magnetic forces or pressure, but through gyration; the entire device gyrates in phase with the payload's 'orbit' through the tube. When the tube is a ring, the gyration must accelerate to increase the acceleration of the payload; the spiral tube has the benefit that the gyration speed may remain constant, although it involves a lot more tubing. These devices usually use magnetic levitation to reduce friction.
It is also possible to make modifications that will add acceleration after launch, to allow for less dramatic launch velocities. One that is often suggested is an 'airspike'3 system that would be mounted on the nose of the ship to create a thermal shockwave ahead of the vehicle, frying a path through the atmosphere using a laser. This greatly reduces the drag caused by the bow shock, and might effectively create all the nose-cone that a ship needs, allowing for lighter ships and a wider range of designs.
Another modification is the possibility that the ship might have some sort of thruster on board to provide an extra push. One common suggestion is a laser propulsion system, where a payload, often solid metal or ice, is super-heated by a laser on the ground, causing it to vaporize and producing thrust. The space fountain is a slightly low-tech version of this, in which some sort of projectile (usually iron) is fired up at the ship, caught, and fired back down. The ship gains momentum both upon catching the projectile and upon firing it back towards Earth.
While the specifics vary from design to design, mass launchers would tend to be much more fuel and cost efficient than chemical rockets, and would be easier and cheaper to set up than a space elevator. They do (or may) have limits on payload size and the rapid acceleration may prevent them from ever being built to transport humans, but they may be our best bet for getting lots of payloads into space in an economically feasible fashion, providing a bridge between the inefficient chemical rockets of today and the economic infrastructure required to justify the immense start-up cost of a space elevator.
A mass launcher of the mass driver/electromagnetic sort could also be used in reverse; a launcher mounted in a spaceship, throwing payloads at high speed out the back of the ship, would act as a propulsion system. This is essentially what an ion drive does on a microscopic scale. This sort of drive might be most practical for an asteroid that needs to be accelerated and hollowed out as a interstellar ship, as it is generally impractical for ships to carry around ponderous loads of mass only to spit it out, bullet-like, back towards their home port4.
1. The original design was actually found in Jules Verne's From the Earth to the Moon, in which he launches his moonship in a moon gun. However, that design was even less practical that the later mass launchers. Superguns are no longer considered a viable method of space launch.
2. They want to use an angled, submerged tube between 400 meters and 1,100 meters long to accelerate the payload with pressurized gas. They expect to be able to launch payloads of 100 lb to 1,000 lbs or more.
3. The Trident missile uses a basic form of this technology, a mechanical structure that disrupts air flow at super-sonic speeds to make a more efficiently-shaped bow shock; this is called a drag-reducing aerospike. Airspikes using lasers are still theoretical.
4. This type of drive might be somewhat practical if the reaction mass was launched in dust-like particles, or if some form of gas was used. While there has not, as far as I am aware, been much development of this sort of technology, the term 'mass launcher' would probably be replaced with a more specific term.
Wikipedia: Mass Driver
The Living Universe Foundation
The Millennial Project 2.0 Wiki: Bifrost
QuickLaunch Hydrogen Gun
Slingatron: NASA technical report
Slingatron with magnetic bearing