Could the solution be "don't actually tie it to the journal"? Have the journal treat itself more as an implementation detail than as an end in itself; you didn't publish a paper in the Journal of Lasers, you just published a paper about lasers.

This is still a mental shift for academics to get over, but I wonder if it's a more feasible one.