2 ms·
Even in science there is too little due diligence. I'm nearly done a PhD in mechanical engineering. Recently I was writing the review section of my dissertatio
by btrettel 7y ago
Even in science there is too little due diligence.
I'm nearly done a PhD in mechanical engineering. Recently I was writing the review section of my dissertation and I decided to check one of the most used results in my field, the most common empirical equation for the boundary of the "atomization" regime in the breakup of liquid jets into droplets. This dates back to the late 1970s. Turns out that it's off by a factor of about 4 due to a math error. Out of the hundreds of people to examine this result, I think I'm the only one to notice the error. I spent a fair amount of time checking the literature to make sure I'm not crazy but seems no one noticed the problem.
A result being highly cited is no guarantee that it's correct.
- mrguyorama 7y agoIs that "off by a factor of 4" result still "useful"? Can you still create products/devices/things/science with that incorrect result? Even epicycles still allowed reduced-accuracy prediction of the movement of heavenly bodies
- btrettel 7y agoYes and no. The result was a factor of 4 too high (edit: 3.3, to be more precise), which is conservative for certain applications. But for those applications I'd argue that you are typically confident that your device is in the atomization regime a priori and don't need to check. But it's extremely inaccurate for my application, where you don't want your device to be in the atomization regime. In fact, I only started examining this in more detail because my own experiments clearly contradicted the textbooks, review articles, etc. If someone blindly applied this result then they'd potentially make a bad device.