Recently, room-temperature crystal structures of SO2F- in its K+ and Rb+ salts were published in Z. Anorg. Allg. Chem. 1999, 625, 385 and claimed to represent the first reliable geometries for SO2F-. However, their almost identical S-O and S-F bond lengths and O-S-O and O-S-F bond angles are in sharp contrast to the results from theoretical calculations. To clarify this discrepancy, the new [(CH3)(2)N](3)SO+ and the known [N(CH3)(4)(+)], [(CH3)(2)N](3)S+, and K+ salts of SO2F- were prepared and their crystal structures studied at low temperatures. Furthermore, the results from previous RHF and MP2 calculations were confirmed at the RHF, B3LYP, and CCSD(T) levels of theory using different basis sets. It is shown that all the SO2F- salts studied so far exhibit varying degrees of oxygen/fluorine and, in some cases, oxygen-site disorders, with [(CH3)(2)N](3)SO+SO2F- at 113 K showing the least disorder with r(S-F) - r(S-O) = 17 pm and angle (O-S-O) angle (F-S-O) = 6 degrees. Refinement of the disorder occupancy factors and extrapolation of the observed bond distances for zero disorder resulted in a geometry very close to that predicted by theory. The correctness of the theoretical predictions for SO2F- is further supported by the good agreement between the calculated and the experimentally observed vibrational frequencies and their comparison with those of isoelectronic ClO2F. A normal coordinate analysis of SO2F- confirms the weakness of the S-F bond with a stretching force constant of only 1.63 mdyn/Angstrom and shows that there is no highly characteristic S-F stretching mode. The S-F stretch strongly couples with the SO2 deformation modes and is concentrated in the two lowest a' frequencies.