Sunday, February 12, 2012
Unicorns, Fairies, and Squatches
Over the past few weeks I have seen numerous references to use cases accompanied by questions.
Why do your tests not cover every use case?
Why would you not test that?
Why do you test like that?
My inner ear rings with the thoughts of “Perfect Software” and “Complete Coverage”.
It is pretty easy to address the concept of perfect software by just handing someone Gerald Weinberg’s great book. Let me quickly simplify the great message Mr. Weinberg presents in this book: “THERE IS NO SUCH THING AS PERFECT SOFTWARE!” Please accept my apologies for the cyber shouting.
Matt Heusser and Peter Walen did a great job at STP Conference in Dallas debunking the concept of “Complete Testing”, but you had to be there. For everyone not in attendance of this wonderful and interact presentation let me quickly share the message: “THERE IS NO SUCH THING AS COMPLETE TESTING!”
While I pondered why so many people still believe in unicorns, fairies, and squatches, I came to another conclusion. Because of the belief in Unicorns, Fairies, and Squatches, testers realistically only get to actually test software 25% – 40% of their working hours. The remainder of the time is spent chasing unicorns, fairies, and squatches. I think this may factor in well with respect to another great book which I am still reading, "How to Reduce the Cost of Software Testing." I need to calculate how much it costs to chase unicorns, fairies, and squatches.
I am so glad that with good testing technique in that limited testing time we typically find most critical defects. I think it is truly OK to let the smaller fairies go. Capturing the Squatches is the most important task at hand!
Do you believe in unicorns, fairies, and squatches?