Keynote Mutation: The Early Days
The start of the testing method that we now call mutation was back when I was a graduate student at CMU. My initial idea was to use the same method that was used in hardware. Hardware testing then (and now) was based on the idea of “stuck-at-faults”. A test set was considered adequate provided it could distinguish between all stuck-atfaults. My idea was then to replace stuck-at-faults by simple changes to the program. One basis of this was the thesis of Young. He showed that many errors in programs were quite simple.
I actually did a thesis on a totally different topic. After arriving at Yale I began to work on mutation once Rich DeMillo joined in. Without Rich mutation would have remained an “idea”. Once Rich was involved we began to move quickly. First, we did some toy experiments. Fred Sayward and others were critical. Then, Rich and his team took over and began to build a series of mutation systems.