When I search for "PTC's in parallel", I find a number of forum posts that say it is a bad idea but none of them say why in depth. I would have though that one would start to open, transferring fault current to the other which would hasten it to go. Is it a bad idea because it is unsafe or is it just an inaccurate method of breaking fault current.
That's exactly the problem (or at least I should imagine it to be so).
The resistance doesn't increase very fast (it's quite slow, on the order of 100ms!), and during that time, the remaining fuses in turn experience far more than their ratings.
For series parts, overvoltage would lead to internal sparking and breakdown.
For parallel parts, overcurrent would lead to hot spots and damage.
I would think parallel would be less bad than series, but still not recommended. (FWIW, conventional fuses are okay in parallel. Ideally, they should be matched resistances. Each fuse must be rated for the full fault current capability of the system, whatever that happens to be. For automotive, this is around 1kA; residential low voltage (120/240V), under 10kA; for industrial 400/480V, 100kA or more; and so on.)
Have you considered using a protected switch, or a hot-swap controller, something like that?
Tim