unlike ^, it does not calculate parity, but whether exactly one operand is TRUE and does so in a short-circuit way.
That seems different to Simon's suggestion, right? So the prior art for using the operator token is different again (although not as different as in D).