Oh great point! Yeah, the problem I think in most of these type of large systems migration choices is that they view only two discrete states, the present state and the target state -- and then try and build and plan for all contingencies. Some of the smarter ones try and define the target state primed to become the next present state with an eye towards a new target state, but in practice I don't think this style of thinking works very well.
In the Soviet example, they were looking at the present state of Japanese manufacturing capability and shooting for that as a target, while the Japanese were improving it every minute of every day. There is no goal in Kaizen, all that matters is the process of improvement. It's a very Zen way of looking at the world that I think thankfully is finding it's way into more modern principles of iterative development (it's amazing how ever-present Kaizen is in Western management training -- but I never got the impression that any of the texts really "got it", instead it's lots of discussion of studying it and trying to figure out how to adapt these two state processes to Kaizen principles without ever understanding the continuous nature of it).
