I use circumstantial evidence frequently when trying to debug software problems. I've gotten particularly good at guessing the cause of a graphics bug just by eyeballing the crazy garbage on the screen.
It doesn't always work. If I can make intuitive leaps five or six steps ahead of what I can prove, I can often diagnose and fix bugs much faster than with a systematic approach to isolating the trouble--of course, I have to do some experiments to verify that I made the right guess. But every so often, I get burned by this approach, because I made a wrong assumption about how the code works or because a coincidence mimicked what I thought was significant. Then, I have to back up and proceed by more methodical baby steps, verifying every step in my deduction--and of course the whole process takes longer than if I had started with this approach in the first place. So there's an art to finding the optimum mix of provable deduction and intuitive induction from circumstantial evidence.
![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png)
timeasmymeasure
no subject
Date: 2007-02-28 03:38 am (UTC)It doesn't always work. If I can make intuitive leaps five or six steps ahead of what I can prove, I can often diagnose and fix bugs much faster than with a systematic approach to isolating the trouble--of course, I have to do some experiments to verify that I made the right guess. But every so often, I get burned by this approach, because I made a wrong assumption about how the code works or because a coincidence mimicked what I thought was significant. Then, I have to back up and proceed by more methodical baby steps, verifying every step in my deduction--and of course the whole process takes longer than if I had started with this approach in the first place. So there's an art to finding the optimum mix of provable deduction and intuitive induction from circumstantial evidence.