GIGO Revisited

Last week I ordered a few small parts from a large, well known supplier. I had been thinking about ordering these for a while but had been putting it off. Then they sent me an email coupon for free shipping and it got me off of the fence and I ordered the parts. The total of the order was around $250. The online catalog indicated that the items "Usually Ship the Same Day". After a few days I hadn't received an email telling me that the order had shipped so I got online to check the status. When I opened my order summary, I was a bit surprised with what I found:

Wow! That is a bit excessive. For that price Scotty should have been beamed the parts to me instantaneously. After a quick phone call to Customer Service the error was fixed and things are back to the way they should be. 

In this case the error was harmless. The company didn't try to charge me $10,000,000 dollars and essentially it was just a typo on a website. But what if this data field were controlling something important?

Way back I wrote about GIGO and blindly trusting the output we get from computers. This glitch illustrates how the computer systems just do what they are told to do by their programmers. If we tell them to do the wrong thing, they will gladly do it, even if it results in disaster. In computer simulations if we put in the wrong boundary conditions, use poor meshing techniques, or even stop the iterations before the solution is converged, we risk basing our designs on flawed data. Any computer simulation must be compared to some benchmark or undergo some sort of "sanity" test. Without these checks, the risk of failures increase dramatically.