VLBI measurements of the size of SN 1979C in M100 (NGC 4321) in the Virgo Cluster, from t = 3.7 yr after the explosion, show an expansion proportional tot(m), which is, with m = 0.95 +/- 0.03, almost consistent with being free for 22 years. The last size measurement, at t = 22 yr, may indicate for the first time a change of the expansion of the supernova and suggests, as an alternative, free expansion proportional tot(m) with m = 1.00 +/- 0.05 up to t(b) similar to 17 yr followed by marginally significant deceleration with m = 0.74 +/- 0.17. The possible deceleration could be weaker within the errors for t(b) < 17 yr or stronger for t(b) > 17 yr. With the assumption for the density profile of the circumstellar medium (CSM) of rho(CSM) proportional to r(-s), we derive a model-dependent value of s = 1.95(-0.05)(+0.10) up to a distance from the progenitor, r = r(b), that corresponds to t(b) similar to 17 yr, which changes to s less than or similar to 1.5 for r > r(b). For a kinetic energy of the shocked ejecta and CSM shells of E-kin = 10(51) ergs, our results require a mass loss to wind velocity ratio for the progenitor of (M)over dot(w)/w similar to 1 x 10(-5) M-circle dot yr(-1) per w = 10 km s(-1), an order of magnitude smaller than estimated from radio light-curve fitting. The swept-up mass at t = 22 yr is then M-sw = 0.3 M-circle dot and the inferred mass of the shocked ejecta Mshock-ej = 2(-1)(+4) M-circle dot. Our last observations give an image of a barely resolved source with a first hint on the structure of the supernova, consistent with being circular within 9% and possibly center filled. The expanding shock front method (ESM) of combining the transverse radio expansion velocities with the radial optical velocities gives direct distance estimates to M100, with standard errors of D = 16.5 +/- 2.5 to 19.8 +/- 3.0 Mpc, depending on whether the supernova has a bright center or is a shell without such a center. These estimates are comparable with those from Cepheid observations (e. g., 16.1 and 15.2 Mpc).