Programmers as mystery writers

As we try to find all the bugs in our program before we release it (at the end of this month!), we are finding that the problems are harder and harder to nail down. Also, they are getting tougher to differentiate. When something goes wrong, is it another case of the bug we've been looking at or is it a new bug? Who should look at this bug? It could be a problem with the database or it could be a problem with the user interface. Or perhaps they are interacting and it involves both of them.
I remarked that sometimes it seems that we're not writing software but mystery novels. A real Who Done It!

We could sometimes use Nick and Nora Charles or even Miss Marple or even Sherlock Holmes to help us track down these problems and figure out who was responsible. In our case, though, we don't want to assess blame but find the person who has the best chance of knowing how to fix the problem.

Then the fellow in charge of testing said maybe it's more like the movies where a little kid tries to tell his parents that there is a monster under his bed. The parents come in to look and find nothing. They leave after tucking the child back into bed and turning off the light. Then the ominous music starts and the program crashes yet again.