Dead Salmon Testing

Earlier this month, there was chatter about the validity of fMRI results that referenced a 2009 study where brain activity was found in a dead salmon. When looking at functionality, it can be tempting to stop at demonstrating that it works.

Image Credit: Jeremy Keith
https://www.flickr.com/photos/adactio/11685017/

But does it fail when it's supposed to fail?

Are you getting expected results only due to coincidence?

What happens when you give your program the equivalent of a dead salmon?

These are the questions that make the testing process take longer that it would seem on the surface.  Showing that "it works" is only a small part of the story.