Agree on the other advice, but how can you claim that a two pole disconnect is only slightly safer than a single pole on an ungrounded DC system? Neither pole is guaranteed to have any relationship with ground potential. What if the design happens to have both poles riding with AC? I think that is possible with NA electricity since 240V will has both poles alternating relative to ground potential.
Now a SCC, ungrounded system, going into a grounded battery, maybe it will have a stronger chance to happen to have one conductor close to ground. But this is not part of the contract in the system type specification (grounded vs ungrounded).
It's not clear what your position is. Are you saying my statement is misleading or false because a two pole is significantly safer than a single pole, or are you suggesting that a two pole is no safer than a single pole?
In my worst case analysis the entire circuit can be boiled down into two cases:
1. Both positive and negative conductors of the source and load are floating or undefined with respect to ground
2. One or the other of the positive and negative conductors of the source or the load is connected to ground
If the circuit is properly wired, and no other fault exists, then in BOTH cases a single pole (switching the ungrounded side in case 2, switching either side in case 1) and a two pole are equivalent. No current will or can flow through the load side conductors if they are connected to ground (through a user, stray wire, tool, etc).
The "slightly safer" looks at possible faults that may occur not in the switch, but elsewhere in the circuit. For instance if one of the two conductors on the source side connects to ground then there are a number of problems that could arise, but since we are focusing on the switch and its protection there are only a few types of faults that would make the difference between a single pole and two pole significant:
In case 1 if either of the source conductors becomes grounded (worn insulation, faulty BMS, leaking electrolyte forming a conductive path, etc, etc) then in a double pole switch there is no path between its output terminals and ground, so there is no shock or fire hazard created by the choice of switch.
If a single pole switch is chosen in this case, though, and if a fault occurs connecting the source that would normally have been switched to ground on the source side, then the switch is ineffective - the unswitched side is now at a different potential relative to ground and the conductors after the switch on the load side are still a shock or fire hazard.
In case two if the source is properly grounded on one conductor, then a single switch on the ungrounded side is only worse than a two pole switch if at least two other failures occur - the grounded side becomes disconnected, and the ungrounded side connects to ground. This is much less likely than a single point of failure required in an ungrounded system to cause the two pole switch to be preferred.
So I still stand by my statement that a two pole switch is slightly safer than a single pole switch. The slightly refers to the fact that another failure must occur for the two pole to be more effective than the single pole. If a system experiences no failures or user errors, then both are equivalent in terms of fire and shock hazards. If a system experiences a single failure, then a two pole switch has some advantages over a single pole switch. If a system is suffering from multiple failures then a two pole has some advantages over a single pole.
But one or more failures or errors must exist elsewhere in the system for this additional designed-in safety factor to become the protective point.
To your second point, if the batteries, which are both a source and load, are connected to another device that is both source and load, such as a combined charger/inverter, then both devices must be separately fused (and switched if the in between conductor is required to be discharged at any point, such as for maintenance) as close to each device as reasonable. If a cable failure occurs between the two (metal cuts through and touches both positive and negative conductors is the trivial example) and you only have a breaker/fuse on one end, then the other end might turn itself into a source and create a fire/shock hazard. Same thing for parallel battery banks - if wire exists between them that is outside their enclosure, then the wire must be fused twice since either end can supply enough current to cause a fire hazard. On most cases each battery pack has an internal fuse/breaker, so this requirement is met, but charge controllers are often not fused internally (which is stupid, IMO, but they do have overcurrent protection provided electronically so they do have some amount of overcurrent protection - just not as dependable as a fuse or breaker)
But the rule of thumb is that all sources must be fused at the source.