Yes, obviously you could give the program the goal of "Stack blocks in order A:B:C" but that's not the point of the conundrum here.
Imagine they are 30 different goals all of which need to be aligned in slightly odd ways to achieve the ultimate task. You're having to prioritize pieces in non-linear fashion to slowly build up the base as you're solving individual steps that seem entirely tangential to the overarching goal. You might have to take 4 or 5 steps back sometimes to achieve 1 step forward.
That's the problem here. It's not this simple 3 block solution: that's only used to get you to wrap your head around the simplest possible complexity.
Re:Sussman anomaly
by hood_syntax on Monday January 15, 2018 @12:15PM
(#16151810)
I feel like the goals are framed incorrectly relative to the context. What's most important is not that B is above C, and A is above B, but that C is on the bottom. If framed that way, you can work backwards from the base case.
That feels trivially like saying: "If you can identify the crucial part of the challenge then it's easy."
The problem then becomes one of identifying the crucial part of the challenge. If framed that way, your comment doesn't really help.
This seems similar to how people tend to focus on short term goals at the expense of longer term ones. Similarly, it's hacking vs architecture in software. You can fail by doing to much of either one.