The big difference between DC and AC is that AC throws you off so you do not remain in contact with the source, DC holds on to you.
The reason for big rail gaps on the underground is to prevent trains bridging sections and so livening up areas where there should be no current.
The reference was relating to DC causing muscles to remain tensed and so holding on to the power source.
That would be a gross over-simplification, you would need to consider the circumstance each time, since repeated spasm caused by AC could be more of a problem in some shock scenarios, where the DC spasm can cause you to be 'held on', similarly it can throw you off, the same is true for AC.
Simply saying DC is more dangerous than AC because it holds on where AC doesn't holds very little truth.
I can say for a fact that having been shocked by both AC and DC at varying voltages, the opposite has usually been true, when receiving an LV AC shock, it has been difficult to pull away from, where as an LV DC shock caused an immediate spasm, tensing the muscle and causing it to pull away from the live terminals.
It all depends on which muscle contracts, and where you're in contact with it, and the state of the muscles when you come into contact.
Either way, my advice would be to simply avoid coming into contact with anything live at all. Aside from some very special circumstances (battery connections being one), it is never a requirement to work live, and indeed there are the electricity at work regulations of the health and safety act that consider this specifically.
And on the topic of bridging,
Bridging has bigger issues than livening up a dead section, the issue arising from the likelyhood of a current sink being present in the dead section.
Lets say that a CRID (Current Rail Indicator Device) and SCD (Short Circuiting Device) have been placed down at either end of the dead section, and a train bridges into it, a full fault current will be drawn down from the local sub station, assuming that it is a double end fed section of course, now if we're near the sub station, this will be of an order of around 10,000A, with the 935mm^2 cable from the TxR (Transformer-Rectifier) feeding the track, lets assume new-ish CCR/ULLCR (Composite / Ultra Low Loss Conductor Rail), this would cause negligible loss compared with the next part of the circuit.
Shoegear drop leads are usually realised by 95mm^2 cable, or 150mm^2 on newer stocks, lets assume an older stock.
This will have 10 times the volt drop compared with the other feeders, the resistance of this being around 0.2mOhm/m. With around 20m of this cable between the 3 way box and the main fuse box, there's 4mOhm per pole of the circuit, so lets now come to our power calculation of I^2.R, 10,000A, squared, multiplied by 4mOhm, so there will be 400kW of power to disipate within the conduits in which this cabling is held. Now consider that the protective devices can take up to 5 seconds to trip, and can be auto re-closed multiple times, the insulation is very likely to have failed and case a flashover on the live voltage circuits of said train.