At the beginning of the week, I showed 
you this picture. 
And I talked about how the structure of 
this part of the course is really 
going to be defined. 
By the cross product of recipes, like how 
to design functions and how to design 
world programs, with data definitions 
with different structure. 
During the week we saw how to design 
simple atomic data definitions, interval 
data definitions, enumerations, and 
itemizations. 
And then we designed one function using 
each of those data definitions. 
We basically completed the red box in 
this week. 
Next week we'll do the green box. 
That point that you can think about the 
organization of the first part of the 
course in terms of a cross product 
between the form of the data and the how 
to design functions recipe. 
That point's important but there's a much 
more important point that underlies that. 
Think about the examples we've seen this 
week. 
Think about the city name example and the 
best function. 
The inherent structure of the domain 
information in that case was atomic. 
So we represented it using simple atomic 
data. 
And that meant the template had a certain 
form, dot, dot, dot x. 
And that told us something about the 
function and something about the test as 
well. 
And that happened with all of the data 
definitions. 
With seednum, the inherent structure of 
the information is a contiguous range of 
numbers. 
So in that case, we use an interval. 
That formed the template a certain way, 
it formed the body of the function a 
certain way, and it told us something 
about the test, in that case, that there 
should be three of them. 
Every example we saw this week, it was 
the same phenomenon. 
Which is that once we identify the 
structure of the information, the 
inherent structure of the information, 
that gives us the structure of the data 
used to represent it. 
Which gives us the structure of the 
template. 
Which gives us the structure of the 
function and also we get guidance about 
the tests for that function. 
So identifying the structure of the 
information is really a key step or maybe 
the key step in the program design. 
At least for many kinds of programs, the 
ones we're going to look at for the first 
part of this course. 
As data definitions get more 
sophisticated, what you're going to see 
is that there are different choices you 
can make about the data that you'll use 
to represent information. 
And by making different choices, you 
actually have a lot of influence on the 
rest of the program. 
Because of this fact that the structure 
of the data determines the structure of 
the template, determines the structure of 
the functions. 
So, designing data is a very high point 
of leverage for designing programs. 
There are some programs for which that's 
less true but for very many programs, 
this approach that we're taking in this 
first part of the course, this data 
driven approach, does a very good job of 
producing the structure of large parts of 
the program. 
What we'll see later in the course is 
some examples where the control structure 
becomes dominant and what we'll learn in 
that part of the course is how to blend 
templates based with control structure 
together with templates based on type 
comments or templates based on a form of 
data to produce the dominant structure of 
our program. 
But even in those programs, the form of 
the information determines the form of 
the data and the form of the data 
determines the form of many of the 
functions that operate on the data. 
So identifying the form of the 
information remains a key step in the 
design of programs. 
That's going to be true throughout the 
course and throughout all programs you 
write. 
So have fun with the homework 
assignments, and I'll see you next week.