The results seem very useful when reasoning about coordination in programs, but I'm questioning how practically useful these results are when applied as methods.
In the GC example, for instance, you do need to find the garbage nodes. Any trivial way of re-formulating that search in a monotonic way would likely destroy performance. My guess is that it's no silver bullet to solving problems in distributed systems.
The results seem very useful when reasoning about coordination in programs, but I'm questioning how practically useful these results are when applied as methods.
In the GC example, for instance, you do need to find the garbage nodes. Any trivial way of re-formulating that search in a monotonic way would likely destroy performance. My guess is that it's no silver bullet to solving problems in distributed systems.