Yeah it would take forever to close for a glow bottle type starter but have you tried connecting a 400v cap in series with the starter it may work because of the inductor in series with the cap to boost the voltage to 230v.
This won't help to trigger the glowbottle starter, as you have to first allow the current to flow in order to get the elevated voltage.
And for the starter internal design:
I think it is matter of manufacturer's preference, if only one electrode or both are bi-metal, it is usually based on the total cost (both options work well): The bi-metal material is more costly (instead of single foil you have to weld two different on top of each other first), while if both are bi-metal's, the machine have to deal with only single type of component (and load it on both sides), so the operation and logistic costs are lower.
The 100/200V difference is mainly in the gas fill pressure and composition and with glow discharge at short distances does not depend as much on the actual electrode distance (there is no room for anode column anyway)...
@Ash: This happen, when the LC is close to it's resonance. But due to higher current the ballast would saturate, it's inductance drop, detune the LC, so it's overall impedance rise and so limit the current. In real life in such configuration is able to maintain the current in wide range of mains voltage, US CWA/CWI ballasts are based on this mechanism (the saturating inductance is there formed by the saturable magnetic shunt between primary and secondary)