From: Stefan Schmiedl <s@...>
Date: 2001-12-08T05:17:32+09:00
Subject: [ruby-talk:27850] Re: John Roth dolt (  Re: A challenge to proponents of Unit Testing. )

lrm@provida.no (2001-12-08 03:35):

> In comp.object John Roth <johnroth@ameritech.net> wrote:
> 
> > Let's think about this for a moment. 5 test cases appears to
> > be adequate for a black box test. You need one for equilateral,
> > three for isosceles and one for scalene. If you're deriving
> > more test cases, you're doing an open box test, based on the
> > actual implementation (or on some set of assumptions about
> > frequent defects, or some such.)
> 
> How about bad data? Won't you need tests to ensure that the code can
> handle bad (zero or negative) values for any combination of the sides?
> Isn't that 7 more tests (illegal data for: {a, b, c, a & b, a & c, b &
> c, a & b & c })?

The user story (msg 27796) said:

"What tests would you write for a function that takes 3 numbers (the
side lengths of a triangle) , and returns whether the triangle is
equilateral, scalene or isoceles."

So I *know* that the input are valid.

Input validation would be done in another method, anyway.
Right now we are focusing *only* on deciding what kind of triangle
we have.

If you check that the numbers make a valid triangle, the Gods of
Refactoring will find ye, and lo, show no mercy on thy code.

BTW: a solution without if could look like

def classifyTriangle(*sides)
  ["equilateral", "isoceles", "scalene"][sides.uniq.size - 1]
end

Now, this raises another question. Given that you wrote the method
this way ... what could possibly break in this method? I'm not talking
about callers here: If a caller sends two or four sides instead of 
three, it's his fault.

s.

-- 
Stefan Schmiedl
EDV-Beratung, Programmierung, Schulung
Loreleystr. 5, 94315 Straubing, Germany
Tel. (0 94 21) 74 01 06
Public Key: http://xss.de/stefan.public

shhhh ... I can't hear my code!