The hot vapor bypass of variable area condenser (elevated) is pretty reliable, but the dynamic performance can be slow. If you go this route I recommend installing instrumentation on the exchanger level. This will allow you to know when you are about to run out of area and otherwise diagnose pressure problems. We have at least 20 of these at this site, but unfortunately not all have the level indication that I suggest. I don't remember all the flavors Sloley mentions, but the one with a control valve at the outlet (drain) of the condenser is the most reliable.
In the reflux drum you cannot count on the "thin crust" of saturated liquid above the bulk subcooled liquid to keep your hot vapor bypass size down. The hazard of trusting to that non-equilibrium phenomenon is that you may find your hot vapor bypass is too small, or that a small level or flow upset will knock down the tower pressure. A few years ago I worked start-up on plant with a tower that used a 2" hot vapor bypass. Last visit I saw that it had been changed out to 8".
You will need a vent in any case for start-up- you should automate this with a valve (HIC if not a PIC). In spite of what Andrew says, I have seen many towers running as a net vapor generator with only the disolved N2 in the feed coming from storage (first tower in train only). You can't really count on this, but if it works then take advantage of it by using your vent for some nice pressure control of the reflux drum.
best wishes,
sshep