Compile Each Concept
We've all been there:
Student: Teacher, I need help Teacher (comes over) Student (shows screen listing three bazillion errors)
The student has just written pages of code and finally decided to try to run it only to end up with pages of errors.
Error messages can at times be hard to read for beginners but to see and truth be told, they frequently don't even read them but over the years I've developed a practice that I've found helpful as a software developer and if students adopt the same practice it can save them a lot of time and effort.
The idea is very simple.
Compile and test one concept at a time.
It might seem silly, but if I'm writing a a program, my first compile might be code that looks like this:
#include <iostream>
using std::cout;
using std::endl;
int main()
{
return 0;
}
or
public class HelloWorld {
public static void main(String[] args){
}
}
This might seem silly but it really doesn't take any effort. I have a
key sequence to do this under Emacs and if I'm using an interactive
language like Python or Clojure I just have to hit cc cc
This might seem silly but I do it out of muscle memory and it immediately tells me I don't have any syntax errors and my build system works.
Going further, I compile and test every time I code up what I call a concept. What's a concept? Let's look at some code. A student might write something like this to find prime number up to n:
def prime_list(n):
for i in range(2,n):
i_isPrime = True
for j in range(i1,1,1):
if i%j ==0 :
i_isPrime = False
break
if i_isPrime:
print(i)
There's a lot going on there. To me, a concept is

A loop

A complex calculation

a conditional
All of these can have other concepts within.
In the above code, I'd probable write it as follows, adding in tests and print statements throughout the process:
Step 1:
def prime_list(n):
for i in range(2,n):
print(i)
Step 2
def prime_list(n):
for i in range(2,n):
for j in range(i1,1,1):
print(i,j)
Step 3
def prime_list(n):
for i in range(2,n):
for j in range(i1,1,1):
if i%j ==0 :
print(i,"is not prime")
break
if i_isPrime:
print(i)
Step 4
def prime_list(n):
for i in range(2,n):
i_isPrime = True
for j in range(i1,1,1):
if i%j ==0 :
i_isPrime = False
break
print(i,i_isPrime)
Step 5
def prime_list(n):
for i in range(2,n):
i_isPrime = True
for j in range(i1,1,1):
if i%j ==0 :
i_isPrime = False
break
if i_isPrime:
print(i)
It might not play out exactly this way but if not it would be something similar. The idea is that if you test every time you add one concept or construct there are fewer places where you can introduce an error.
If you enter 100 lines before you test there are 100 places where things can go wrong. If you type 10, there are only 10. On top of that, if you've added 100 lines, conecptually you've probably added a lot and the error can be anywhere. While it's not always the case, most of the time, if you just added an if, the problem will be in the if or as a result of the if. Same with a loop or any other construct.
Once you get in the habit, it's easy and doesn't really take any time. A couple of keystrokes to compile and a couple more to run.
All too often students try to write everything at once and it's so rare that it works. If we can get them to develop incrementally they'll be able to write much more complex systems and write them with much less frustration.