There are any numbers of things that will be a matter of taste for each smith, I see no sense in any of the vertical cylindrical designs, but users of them have issues with the horizontals that I favor, so everybody will have their own reasons a rationales for what they prefer. The one feature that cannot yield to alternate rationales or opinion is the fact that corners make cold spots, so not only do I totally agree with what has been said here, I believe the facts also totally support it. For most operations that we do with our steel choices even heating is most critical, with the only exception being localized heating for specific forging operations where one doesn't want to overheat thinner sections that are already forged, but this is dealt with by adjusting our forging techniques ans sequences when making the transition from coal to gas. With no concern for vortex action or even heating many square forge designs simply have the flame coming straight in and directly onto the work- this is a no no for blade forging, avoid any design that has the flame blasting directly onto the side of your blade.
All gas forges should work off reflective or conductive energy, the flame does not do the heating of the blade, the flame heats the walls and either has it reflected to the center (wool line forges) or conducts it to the center (hard refractory lined forges). It should be immediately obvious that a cylinder will send this heat back to the center in the most even fashion. A flame entering on a tangent and creating a vortex around those outer walls will increase this effect even more, thus my strong preference for a design that allows me to pass the steel down the center of that vortex instead of through the side of it.
In short, I know - too late, you can bank on what Thunter and Ed have told you and add my vote to the count.