Flying With Python (Strong versus Weak Typing)

Roy Smith roy at panix.com
Thu Mar 13 14:33:32 CET 2003


"Delaney, Timothy C (Timothy)" <tdelaney at avaya.com> wrote:
> I don't claim to follow the XP way, but I am a strong believer in unit tests.

I've found a few very non-XP uses for unit tests.  For example, 
sometimes they can be used to force people to document their work.

A while back, we embarked on a project to add unit tests to a library of 
utility functions which had been in production use for a while.  We 
found a few bugs, but more importantly (in my mind), it forced people to 
go back and carefully define and document the behavior of each function.
  
There was a function called isValidIpAddress.  Turns out the original 
author had one thing in mind for the definition of "valid", and other 
people had used the function thinking it did something different.  The 
issue had come up at various times in reviews of code that used this 
function, but somehow we always managed to argue ourselves out of really 
figuring out what it was supposed to do.

I wrote a unit test for the function, using my understanding of what 
"valid" meant.  The existing code failed the unit test.  Now we had 
something which was harder to ignore.  Instead of vaguely-remembered 
arguments at code reviews, we now had a "FAILED" message in our face 
that we had to do something about.  Either the code had to be changed to 
pass the unit test I had written, or somebody had to write a precise 
enough spec for what the function was SUPPOSED to do, and then I would 
change the unit test to match the spec.

If you don't know what it's supposed to do, you can't tell if it's 
working or not.




More information about the Python-list mailing list