On a Compact Disc, data are represented by microscopic holes on the disc surface. The data are then read back by a laser. Since the rearranging of holes on a Compact Disc surface requires physical alteration of the material, re-recording a CD causes the medium to degrade quickly. With the MD, Sony aimed for a digital music format that could easily be re-recorded many times over without quality loss. Therefore, Minidisc had to employ a very clever alternative to the land-and-pit scheme of the CD.
On Minidiscs, data are represented by magnetic charges instead of lands and pits. A polarized laser is then used to read the data. In polarized light, all waves are in a single three-dimensional plane. When a polarized beam hits a magnetic surface, the polarity of the light changes on reflection: the lightwaves will be in a different plane.
Exactly how the polarity changes depends on the polarity of the magnetic surface. Thus, a laserbeam reflecting from a positively charged surface will have different polarity than a beam reflecting from a negatively charged surface. If you use special Polaroid 'glasses' that are aligned with, say, the positive reflections, you would only be able to see the light if the reflective surface is positively charged. This way, you can 'see' the polarity of the reflective surface.
In order to use the phenomenon described above to store data on the Minidisc, a technique is required to give very small parts of the Minidisc surface a specific magnetic charge. There are two ways of doing this. Both are based on the fact that any potentially magnetic material will instantly take over a surrounding magnetic charge when the material is heated to its Curie point.
A Minidisc recorder contains a crude electromagnetic coil that can put either a positive or a negative magnetic field around a part of the Minidisc. The disc surface is made of a material that is not easily magnetized. When the recording laser heats a small area on the disc, it will take the magnetic polarity of the coil.
A disc surface can first be charged negatively. Then the coil can be set to produce a positive magnetic field and the laser can be used to heat up all the desired areas on the disc to their Curie point. These areas will then have a positive charge and the data can be represented conveniently.
Another way of doing it is by keeping the laser hot and directing it past all sequential surface areas. The field direction of the magnetic coil can then be modulated above certain areas to produce the desired charge. This method is a bit harder to realise because it requires a more delicate coil. The magnetic field direction of a coil is more difficult to modulate than the amplitude of a laser. One serious benefit is that a surface can be re-recorded in one go. It doesn't need to be reset to an all-negative starting position.