The separation of DNA fragments by gel electrophoresis has been studied extensively over the last two decades. More recently, similar studies have been carried out to characterize the separation achieved by the current capillary array electrophoresis systems and their sieving polymer solutions. In all cases, at least three different mobility regimes have been shown to exist: the Ogston regime when the radius of gyration of the DNA fragment is smaller than the pore size, the reptation regime when the DNA is larger than the pore size but remains in a random coil conformation, and finally the reptation-with-orientation regime where the DNA orients in the field direction and essentially all resolution is lost. Unfortunately, although theory helps us understand the different regimes and how to properly exploit them, we still have no theory-based general equations that would apply to all regimes. Such equations would be especially useful to analyze data, optimize separation systems and interpolate mobilities to estimate unknown molecular sizes. Recently, van Winkle, Beheshti and Rill (Electrophoresis 2002, 23, 15-19) proposed an intriguing empirical formula that seems to adequately fit the mobility of dsDNA fragments across all three regimes. In this paper, I investigate the relation between this empirical formula and the known theories of gel electrophoresis, and I study the dependence of its fitting parameters upon the experimental conditions. Finally, I examine how this equation may need to be modified to capture the more subtle details predicted by fundamental theories of DNA gel electrophoresis.