I got 99 (Fermi) problems…
Ever wanted to calculate the yield of an atomic weapon in terms of equivalent tons of TNT? Would you ever get the opportunity?
Well, Enrico Fermi wanted to do this in 1945 at the first testing of an atomic bomb in the New Mexico desert. As the shock wave passed he dropped small pieces of paper and the fact that they ended up around 2.5 meters (8 feet) from his body allowed him to estimate the yield as around 10 kilo tons equivalent of TNT – immediately without apparent thought. A team of MIT scientists worked for a further 3 weeks and came up with the ‘true’ value of 18.6 KT. So Fermi was right within a factor of 2 with a simple experiment and ‘data treatment’ (maybe a ‘lookup table’ in his head).
This is the original Fermi-type problem – working out something with limited data and with appropriate estimation. Two classic examples are:
- How many piano tuners in Chicago?
- How much is spent on haircuts in the USA in 1 year?
For a number of years I used to ask (from time-to-time) a borrowed (not my original) interview question along the lines of “You’re over the Mariana Trench in a boat with a cannonball in your hand and drop it overboard. How long does the cannonball take to reach the bottom?” The rules that there are no rules except that the response (note: response not answer) is required within 24 hours (not within the interview time). This response time allows the interviewees time to get back to Malvern but also the ability to work on a ‘real’ industrial-type problem where they have access to all forms of data (libraries, books, personal knowledge and contacts, the Internet….) as opposed to a closed-book examination type situation. From our (my?) perspective I could see how the responder would tackle and present their evaluation (note not solution) of the problem. I hoped to excite some enthusiasm similar to that I had as a schoolboy when the Trieste with Jacques Piccard and Don Welsh descended to the bottom of Challenger Deep – the deepest part of the ocean.
We do have some take-home messages here – there is a possible competing technology of sedimentation/settling that we need to be aware of and also the routes to solving a problem where there isn’t a rigid fixed answer but where some estimation and research is needed to justify any stated answer.
So have a think yourself. How would you tackle the Mariana Trench problem? Try the 24 hour response test before viewing: Solving the Mariana Trench Question.
The question won’t be asked again by me….
Take a look at Careers at Malvern for the chance to answer interview questions such as this one!
Attending Pittcon in 2015? Check out our short courses.
More posts from Alan Rawle:
Weighing heavy? It’s a matter of scale