Nigel's Law

I test open source software that scales to thousands of nodes. In the course of my work, I occasionally hear from the project developers that certain test cases are "corner cases" -- that is, the test cases are covering rare or infrequent situations.

Recently I was explaining this developer behavior to a colleague. I explained to him how I was weary of test cases being passed off as "corner cases" that didn't merit much attention. What I told him next he subsequently dubbed "Nigel's Law", so here it is:

In a large enough cluster, there are NO corner cases

The fringe, the boundary, the infrequent, the unlikely -- all occur quickly and often when operating software on large clusters with thousands of nodes.


Want a fun job testing open source? Passionate about software quality? Have great testing and coding skills? I'm hiring! Send your resume to: ndaley at yahoo-inc dot com