Laser cooling and trapping offers the possibility of confining a sample of radioactive atoms in free space. Here, we address the question of how best to take advantage of cold atom properties to perform the observation of as highly forbidden a line as the 6S-7S Cs transition for achieving, in the longer term, Atomic Parity Violation measurements in radioactive alkali isotopes. Another point at issue is whether one might do better with stable, cold atoms than with thermal atoms. To compensate for the large drawback of the small number of atoms available in a trap, one must take advantage of their low velocity. To lengthen the time of interaction with the excitation laser, we suggest choosing a geometry where the laser beam exciting the transition is colinear to a slow, cold atomic beam, either extracted from a trap or prepared by Zeeman slowing. We also suggest a new observable physical quantity manifesting APV, which presents several advantages: specificity, efficiency of detection, possibility of direct calibration by a parity conserving quantity of a similar nature. It is well adapted to a configuration where the cold atomic beam passes through two regions of transverse, crossed electric fields, leading both to differential measurements and to strong reduction of the contributions from the M1-Stark interference signals, potential sources of systematics in APV measurements. Our evaluation of signal to noise ratios shows that with available techniques, measurements of transition amplitudes, important as required tests of Atomic Theory should be possible in 133 Cs with a statistical precision of 10 −3 and probably also in Fr isotopes for production rates of > ∼ 10 6 Fr atoms s −1 . For APV measurements to become realistic, some practical realization of the collimation of the atomic beam as well as multiple passages of the excitation beam matching the atomic beam looks essential.