#Effect of Edge Cases on Coding Practices
An experiment was performed to test the effect of giving thought to edge cases before begininng coding. For more information see details on the experiment and instructions for the control group
##States of the program
To compute a progress to solution metric, we classified each submission to be in one of many states, based on the effect of the code when compiled and tested against various test cases.
###Delete
A – Compile error
B – Runtime error (LOSSY)
C – Nontermination (LOSSY)
D – No effect
E – Neither beginning nor end (unsure)
F – Deletion works at nodes that are neither beginning nor end
G – Only deletion at end works
H – Deletion at beginning doesn’t work
I – Deletion at end doesn’t work
J – Cannot delete from beginning of a non-singleton list
K – Cannot delete anything except beginning of non-singleton list
L – Cannot deleting anything except beginning of non-singleton list (empties list)
M – Works on all lists of length <= 2, but cannot delete from middle of list of length >= 3
N – Can only delete from beginning of non-singleton list (all singleton lists are fine)
O – Deletion of an element from a list of length >=3 deletes the entire list after it
P – Failing one case (deletion of second element of a list of length >= 3
Q – Cannot delete from end of a non-singleton list?
Y – FAILURE (no students reached solution once they entered this state)
Z – SUCCESS