Interview: How would you test this function?

During a programming interview, once you finish writing a function, they often ask you to test it. What are some must-know test cases that apply to most functions one can encounter in an interview?

Thanks!
gromulAsked:
Who is Participating?
 
rstaveleyConnect With a Mentor Commented:
Good buzz words for an interview are: purpose, promise and requirement. I've always known these as the basis for "programming by contract", but there may be some other name for this.

Your code of course has a purpose. Based on the stated purpose, it makes promises, which it must fulfill at the very least (e.g. it will make a pizza within 10 minutes). It can exceed these promises (e.g. make the pizza in 8 minutes), but to achieve its promises, it needs to state its requirements (e.g. pre-heated oven, cheese) and not expect anything better than its documented requirements (e.g. if it only says it requires cheese, it isn't OK for it only to work with Mozzarella cheese).

Unit tests demonstrate that a function fulfills its promises with a range inputs that fulfill its requirements. Thorough testing explores the entire range of requirements (e.g. testing the good pizza code with mouse trap cheese and with a pre-heated over that's absurdly hot). Because it is hard to test thoroughly all possibilities for code that makes few requirements, it is worth narrowing its documented requirements for the sake of testing. It can seem like its "the tail wagging the dog", but code is much more valuable when it can be tested.
0
 
e_tadeuConnect With a Mentor Commented:
The test case will depend on what your function is supposed to do. Normally, the test consists of ensuring some post-conditions. A very simple example: if you must test a function that negates a number, you can test like this:

assert(Negate(-3) == 3);
assert(Negate(10) == -10);

You can also test invariants too.

See:
http://en.wikipedia.org/wiki/Test_case
http://en.wikipedia.org/wiki/Unit_test
0
 
dbkrugerConnect With a Mentor Commented:
For integers you should consider things related to problems with word wrapping. For example, on a 16 bit unsigned short, -1 == 65535, so there might be surprises

common values 0, 1, -1, -32768, 32767

for floating point, issues involve division by zero, sqrt of a negative, overflow, underflow. These get more complicated as the intermediate calculations get more involved.
0
Question has a verified solution.

Are you are experiencing a similar issue? Get a personalized answer when you ask a related question.

Have a better answer? Share it in a comment.

All Courses

From novice to tech pro — start learning today.