As a correction, there is no abstract. Here is a summary:
Unfortunately, my paper contains a mistake. I wrote that "[b]y measuring the difference in redshift between two images, one effectively measures the time delay". However, the difference in redshift is proportional to the to the time delay and to the Hubble constant (my equation 5), thus one cannot use the measured difference in redshift between images of a gravitational-lens system in order to measure the Hubble constant. All other aspects of my paper are unaffected: consistency check on standard cosmology, using an Einstein ring to correct for the change in redshift due to the moving-lens effect, effectively measuring time delays (for a given Hubble constant!) on much longer timescales and/or for non-variable sources, measuring the transverse velocity of the lens, and determing which images in a complex gravitational-lens system correspond to the same source.