Now return to pp 16-17 for comments on the Evans Analogy Program, and then read pp 271-288 in the introduction to Evans' article in the same book. As p 281 makes clear, Evans wrote his program as an experiment in the use of ``good'' ``high-level'' representations of simple pictures. This was then an idea still new to computing and AI - now it's old hat. So the Analogy program is worth knowing because it demonstrates the importance of representation, and why some representations are better than others. By trying to simulate its behaviour, you can get a feel for this.
Evans goes on to describe in detail how his program operates. Apart from the summary on pp 288-290, you'll probably find this account too technical. However, there's an easier one written by Patrick Winston, formerly one of Evan's students at MIT. This is in pp 24-33 of [Artificial Intelligence: Winston]: it includes a discussion of what makes a representation good. So read these pages, and work some examples. How does the representation restrict the class of analogies the program can solve?
One extra point: you should be aware that Winston's account is different
from Evans' own, being simplified to make it easier to understand. The
program was actually more complex (to make it more efficient). However,
the principle, of finding the difference between 
and 
, and then
looking for that figure 
which differs from 
in the same way as
differs from , is the same.