The radon absorption ability and the track etch properties of the polycarbonate material of commercial compact disks make them very useful as sensitive retrospective 222Rn detectors. The basic idea is to remove, after exposure, a surface layer that is thicker than the range of the alpha particles of the 222Rn and 220Rn progenies and to count the electrochemically etched tracks at the corresponding depths (>80 microm). The effects on the response due to differences in pressure, temperature, and humidity have been studied experimentally. The effect of the growing of 210Po after long-term exposures was also estimated. The effect of all listed factors except the temperature is either absent or restricted to maximum--about 10% for the very extreme cases. The variation of the response at 83 microm depth over the temperature interval 15-25 degrees C is +/-12% around the 20 degrees C value. The dependence of the calibration factor on the etched depth beneath the surface was studied at 4 different temperatures within the range expected indoors. The results show that the depth dependence is exponential with the parameters of the exponent also being dependent on the temperature. In practice, using the track density obtained in two or more depths beneath the compact disk's front surface, an a posteriori temperature correction could be made. By this correction it is possible to substantially reduce the bias in the results due to the unknown temperature during exposure. The results imply that by using home stored compact disks long-term retrospective 222Rn measurements could be made with an uncertainty that could be potentially better than 10%. The useful range of the method starts at about 3 Bq m(-3) (for 10 y exposure time) and appears to cover practically the whole range of indoors 222Rn concentrations.