Going so wrong that it has to be right August 11, 2011
Posted by mareserinitatis in engineering, research.Tags: engineering, engineering research, mistakes, research
trackback
The most exciting phrase to hear in science, the one that heralds the new discoveries, is not ‘Eureka!’ (I found it) but ‘That’s funny…’ 
– Isaac Asimov
I’d have to say that one of the most interesting things that’s happened to me at work was a simulation I ran recently. I got this jaw-dropping awesome result. Then I went back to look at my device I’d modeled and discovered I’d modeled it incorrectly.
I was very disappointed because I was really hoping this idea would work, but it didn’t. When I made the corrections, things had still improved, but not nearly as much as when I’d done things wrong.
A week or so later, however, we got the idea of, “What if we could make this device?” And it turns out that we probably can, only using a very non-standard method.
I have to admit that I like to have my work planned out so that I can approach it in a methodical pattern. I like to do this so that I can make sure my results are very consistent, and it’s easier to tease out problems when things don’t look right. (I learned that the hard way.)
This is the first time in my experience, however, where a mistake ended up being a good thing. I wouldn’t mind having more experiences like that.
So have you ever made a mistake only to discover that it wasn’t?
I also blog at Engineer Blogs, home away from home to some of the best engineering blogs.
Yes, it has happened to me several times, even recently — and yes, that is often how the most interesting things are discovered ;-)
We tried to model a two-dimensional system of particles possessing an electric dipole moment, and because all moments would be aligned perpendicularly to the plane, the interaction would be purely repulsive, growing as the inverse cubic power of the distance at short distances.
Now, my postdoc, who was not familiar with the use of my code, ended simulating inadvertently a system in which the interaction, instead of growing, flattened off at short distance, giving rise to very neat results. In fact we were very excited when we saw them… Until, that is, we realized that they were due to a screw-up. The actual system does not show the same interesting behavior — much more boring instead.
The unexpected turn if event came a month later when I was jokingly describing over lunch the screw-up to a colleague working in a slightly different area. Instead of laughing about it, he went “you know, actually …”. As it turns out, something which I thought was unphysical (the flattening off of the potential at sort distances) can in fact happen under specific physical conditions, rendering our “wrong” simulation quite interesting and relevant… Moral of the story, we ended up publishing the “screw-up” in Physical Review Letters. Of course if someone asks me now I tell them that it was my brilliant intuition and deep knowledge of physics…
I have to admit that, up until now, my screw ups have not resulted in much other than wasted time. I certainly wish I could say that it was a result of incredible intuition, though. :-)
[...] always happen. And, in fact, getting results that are ‘off’ is sometimes good as it can lead to new areas of research. However, more often than not, it can also be a result of bad [...]