It thus appears that a long-standing riddle has now been solved. Fast quasi-Alfvénic magnetic reconnection may occur under rather general conditions with a rate rather independent of the particular reconnection physics, both in high- and low-β plasmas. Ironically, the case of stationary resistive MHD, which has been regarded as the most natural framework of reconnection theory, does not allow fast merging. The pecularities of resistive reconnection have been the origin of the long controversy dividing the community into two camps, the adherents of the Petschek model and those of the Sweet-Parker model emphasizing current sheets. Actually, physical conditions for a stationary high-Lundquist number MHD model to apply are rarely satisfied, neither in nature nor in laboratory plasmas. Either the plasma is strongly resistive, which often implies relatively low S or, at large S-value, collisionless effects are more important than resistivity. In addition, the plasma behavior is usually highly nonstationary, sometimes fully turbulent, and such a system allows fast reconnection even in resistive MHD. It is true that the usual quasi-stationary 2D models for collisionless reconnection are also highly idealized, far from real plasma conditions. Real plasmas which tend to exhibit a whole maze of fluctuations, but these do not seem to control the reconnection rate.

A concept which had also polarized the community, the distinction between driven and spontaneous reconnection, has to a good deal lost its significance. It is true that in many cases an external driving agent can be identified.