Has anyone calculated what the cost of doing this correctly once vs the ongoing support/SLA/etc issues of repairing it when it goes boom is? I've gotta believe that for >90% of the situations where diverse routes exist, just being used as dual linear paths, its cheaper in the long term to do it "right" and cut the size of your outside plant crew (assigned to break/fix) by 90%. :)
It is definitely cheaper to "do it right" the first time, but to "keep it right", the operator needs to pay very close attention to each and every circuit groom that they perform, less they end up with a degenerate loop somewhere. Depending on the state of their circuit routing database(s), the exercise of checking for overlap against "the other half" of the ring can anywhere from trivial to impossible. I don't think operators intentionally foul this up, but it's real easy to get wrong, particular in the fallout after accumulating a bunch of different companies fiber plants and circuit route systems and trying to consolidate everything for savings. I'm just providing this in answer to "why can't telcos get this right", as there's no reason to think that grooming activity was at all involved in why this particular carrier got stung... ;-) /John