It's finals week here at UC San Diego, and I can practically see the tumbleweeds making their way across our deserted streets as multitudes of students flee into their caves. The few that can be found out and about more closely resemble zombies than some of California's brightest. There's practically a tradition here that no one is allowed to do anything fun during finals week because they have to cram for finals, and I think it's completely boneheaded.
I'm not saying a bit of review before your tests is a bad thing, but it's hard to learn in one all-nighter what you've failed to learn during ten weeks of classes; if you've done your job during the quarter then your spare time during finals should be spent resting and relaxing. People perform much better when they're happy and rested than when they're stressed and tired. That should be obvious, but students everywhere seem hell-bent on running themselves through a gauntlet right before they have to be at their best for three hours straight. Would Lance Armstrong train hard all night right before a race? I don't think so. Think of your final exams like the most important athletic event you'll ever compete in, and prepare accordingly: train hard for months, and rest up as much possible the day before. It'll make you not suck as much at science, I promise.
Tuesday, June 12, 2012
Wednesday, May 30, 2012
Ah, Singapore
This overpass is part of the international airport in the small powerhouse nation of Singapore. That's some pretty epic engineering, lah.
Show Your Work.
One of my close friends tutors a couple of kids in math, and recently they were complaining to her about needing to show their work. She told them that she had a friend who could solve very difficult differential equations in his head, and that they had to show their work until they were right as often as he was.
There is a general viewpoint in academia that it is necessary to write down work because it helps avoid errors. Students have to show their work on tests so that professors can tell they aren't cheating. Not only do I respectfully disagree, but I believe they are completely missing the point. Encouraging students to avoid doing work in their head is actually doing them a significant disservice; fast mental math is a very valuable skill in the real world. Richard P. Feynman himself was very proud of being able to complete difficult arithmetic computations faster than most abacus users. Especially in engineering, the ability to ballpark estimate performance figures or material requires quickly without needing to reach for pen and paper is a very useful one indeed. Imagine being part of an aircraft design team and having your boss ask "what if we crammed 50 more seats into our new airliner?" You want to be the guy who can come right back with "the extra weight would decrease our range by about 5%." Aside from being a useful skill, doing work in your head does wonders for your mental sharpness and memory. Forcing yourself to keep track of all those variables without writing them down expands your mind and helps you think differently. It will help you remember what you're doing better and make you a better engineer in general. Both professors and students place far to little value on mental math abilities, and it's the students that eventually suffer for it.
That being said, I don't condone never showing work, especially on tests. Writing down every step helps you double check that your mental math is correct; being fast and wrong is worse than useless. More than that though, showing your work is a communications skill. What's the use of being able to do all the analysis in the world if your boss and colleagues can't understand it? Writing down and explaining everything you do is worthwhile because being able to quickly convince people you're right is as valuable a skill as being right fast, if not more so.
So where does that leave us? Do all of your problems completely in your head first, even on tests. Then write down your solutions in terms even a freshman could understand. This will help you get the most from your classes and be the most successful in life. One of my favorite professors put it this way: "There are three kinds of students: the smart ones who will start a problem immediately and slow down when they get a bit stuck, the stupid ones who will have no idea how to do it, and the true geniuses who will write nothing down until they've solved it then write down their whole solution in a minute or two."
There is a general viewpoint in academia that it is necessary to write down work because it helps avoid errors. Students have to show their work on tests so that professors can tell they aren't cheating. Not only do I respectfully disagree, but I believe they are completely missing the point. Encouraging students to avoid doing work in their head is actually doing them a significant disservice; fast mental math is a very valuable skill in the real world. Richard P. Feynman himself was very proud of being able to complete difficult arithmetic computations faster than most abacus users. Especially in engineering, the ability to ballpark estimate performance figures or material requires quickly without needing to reach for pen and paper is a very useful one indeed. Imagine being part of an aircraft design team and having your boss ask "what if we crammed 50 more seats into our new airliner?" You want to be the guy who can come right back with "the extra weight would decrease our range by about 5%." Aside from being a useful skill, doing work in your head does wonders for your mental sharpness and memory. Forcing yourself to keep track of all those variables without writing them down expands your mind and helps you think differently. It will help you remember what you're doing better and make you a better engineer in general. Both professors and students place far to little value on mental math abilities, and it's the students that eventually suffer for it.
That being said, I don't condone never showing work, especially on tests. Writing down every step helps you double check that your mental math is correct; being fast and wrong is worse than useless. More than that though, showing your work is a communications skill. What's the use of being able to do all the analysis in the world if your boss and colleagues can't understand it? Writing down and explaining everything you do is worthwhile because being able to quickly convince people you're right is as valuable a skill as being right fast, if not more so.
So where does that leave us? Do all of your problems completely in your head first, even on tests. Then write down your solutions in terms even a freshman could understand. This will help you get the most from your classes and be the most successful in life. One of my favorite professors put it this way: "There are three kinds of students: the smart ones who will start a problem immediately and slow down when they get a bit stuck, the stupid ones who will have no idea how to do it, and the true geniuses who will write nothing down until they've solved it then write down their whole solution in a minute or two."
Keep it simple, stupid.
I feel like over-complication is a big problem we face in the world today. It's true of exercise plans, tax forms, relationships, and just about everything that is engineered.
My father always used to complain about how engineers are almost never required to learn much optimization, and I'm starting to understand what he meant. We spend countless hours learning to analyze the forces acting on our designs, but very little attention is paid to whether we're doing things in the best way possible.
It is true that engineering can be as much an art as a precise science, however I feel that form should always follow function. A machine had better damned well work reliably before the designers go about embellishing it with bells and whistles. There are several successful foreign carmakers that have made themselves infamous for being evidently unconcerned with making cars that actually run, but I am not a fan of this approach.
Designing for functionality is not a difficult concept: figure out all the functions a machine absolutely must perform, and build something that will reliably knock them all out of the park. It's the execution that trips most people up. How do you build something that gets the job done right every time with minimal development costs and maximum efficiency?
Simplicity. Start with the absolute minimum functional requirements of your project, and think about how to do that with the absolute fewest number of moving parts, actuators, etc. Add complications only as they become necessary; make it work in its most simple form before making 'improvements'. Every added complication or moving part is something else that can get jammed or otherwise fail and something else requiring rigorous development. Make sure everything works well before adding to it, and start simple. Not a very difficult concept, but surprisingly few people seem to want to adhere to it.
My father always used to complain about how engineers are almost never required to learn much optimization, and I'm starting to understand what he meant. We spend countless hours learning to analyze the forces acting on our designs, but very little attention is paid to whether we're doing things in the best way possible.
It is true that engineering can be as much an art as a precise science, however I feel that form should always follow function. A machine had better damned well work reliably before the designers go about embellishing it with bells and whistles. There are several successful foreign carmakers that have made themselves infamous for being evidently unconcerned with making cars that actually run, but I am not a fan of this approach.
Designing for functionality is not a difficult concept: figure out all the functions a machine absolutely must perform, and build something that will reliably knock them all out of the park. It's the execution that trips most people up. How do you build something that gets the job done right every time with minimal development costs and maximum efficiency?
Simplicity. Start with the absolute minimum functional requirements of your project, and think about how to do that with the absolute fewest number of moving parts, actuators, etc. Add complications only as they become necessary; make it work in its most simple form before making 'improvements'. Every added complication or moving part is something else that can get jammed or otherwise fail and something else requiring rigorous development. Make sure everything works well before adding to it, and start simple. Not a very difficult concept, but surprisingly few people seem to want to adhere to it.
Subscribe to:
Posts (Atom)