The next three videos are examples of 
designing a function that consumes 
non-primitive data, data that's defined 
by your own data definition. 
Now, they are HtDF problems so mostly you 
know how to do them. 
Remember, I said the HGDF recipe is 
largely orthogonal to the form of data. 
But it isn't completely orthogonal 
[INAUDIBLE] form of data. 
One real dependence between function 
design and form of data has to do with 
the number of tasks that's appropriate 
for that function. 
So in these videos I'm going to pay 
particular attention to the number of 
test that we're going to use based on the 
form of data. 
 >> I mean aisle-starter.rkt and the 
problem we're asked is to use the SeatNum 
data definition. 
And design a function that produces true 
if the given seat number is on the isle. 
So the first thing I would do is go look 
at the SeatNum data definition. 
Then it reminds me that seat number is a 
natural number from 1 to 32 inclusive, 
and that its seat numbers in a row and 1 
and 32 are the aisle seats. 
So now I can get started on my function. 
I'll quickly do the signature purpose and 
stub. 
Let's see, it's going to consume a 
SeatNum and we're supposed to produce 
true when the seat number is on the 
aisle. 
So that means it's going to produce a 
Boolean. 
And it'll produce false when the seat 
number isn't on the aisle. 
[SOUND] And let's see, produce true if 
given seat number is on the aisle. 
[SOUND] And, let's see the function. 
This is a predicate, so we'll name it 
using the naming convention for 
predicates like this, (aisle? 
sn) is and we'll just produce false for 
the stub. 
So let me run that to make sure the stub 
is well formed than it is. 
Okay. 
Now let's get to the examples. 
Now here is the first place where the 
data definition is really go have 
something to say. 
If we go back to the How to Design Data 
page, and we go to the section on 
Intervals, and we go to the section on 
Guidance on Examples and Tests, it tells 
us that when writing tests for functions 
operating on intervals, we should be sure 
to test closed boundaries as well as mid 
points. 
And always be sure to include enough 
tests to check all other points of 
variance. 
So let's go back to the code. 
I've got an interval with two closed 
endpoints. 
So I know that what I'm going to need are 
three tests. 
One [SOUND] for the first closed end 
point which is 1. 
1 for the middle, which is, you know, 
middle doesn't have to be the exact 
middle. 
Something like the middle. 
Well call it 16 and 1 for the closed 
upper end point which is 32. 
Now what are these going to be? 
Well, I'm supposed to produce true if the 
given seat number is on the aisle. 
Lets see. 
This one is on the aisle. 
This one is not. 
And this one is. 
Notice how much information the data 
definition gave me about generating the 
test. 
Data definitions are always going to be 
helpful this way. 
They're not always going to make it quite 
this straightforward but data 
definitions, the structure of the data 
definition, the structure of the data 
tells us a lot about the test we need. 
So, let me run that. 
Good, they're all running. 
Now for the template, what I'm going to 
do is, I'm first going to come in off the 
stub. 
I'm going to say use template from 
SeatNum and I'll just leave that note or 
copy the template down. 
You know, use it like that. 
I have to rename it to the aisle, that's 
the name of this function. 
And let's see. 
I'm consuming a seat number, which is a 
natural from 1 to 32. 
And I need to produce true if it's an 
aisle seat. 
That means, I need to produce true if 
SeatNum is 1. 
Also SeatNum is 32, and I just need to 
order those two together. 
Let's try that. 
All three tests passed. 
So, there you go. 
That's designing a function that consumes 
an interval. 
And the big piece of help we got from 
knowing it's an interval is, well, of 
course we got the template, but we also 
got the structure of the tests. 
In this case, it's an interval with two 
closed ends. 
So, we have a test for each of the ends 
and a test for somewhere in the middle.