Quote:
|
Originally Posted by KenWittlief
The temptation to jump into a project at one point, and fly by the seat of your pants is overwhelming. For most new engineers, once you have done than, and gotten burned by your own ego, then you really understand the whole design cycle, and why each part is absolutely necessary.
|
I've had the same experience with our team's students: they want to jump in immediately and do everything (and have it work right the first time). Every year it's a long effort teaching them the virtues of incremental development.
Typical conversation:
Student: Why are we testing trig functions on the bench? Why don't we just write the code and run it on the bot?
Mentor: Because if we test the trig functions (and the gyro, and the camera, and ...) here, we'll know the math is right, so when the bot spirals around randomly instead of doing the right thing, we will have a much better idea of what went wrong.
Student: But I want to run the bot now.... I have my completely untested follow-the-camera-with-PID-control-gyro-and-accelerometer-code that I want to try out.
Mentor: You must first learn patience...
Student (impatiently): How long is that going to take?
However, it's working now. All the nuts and bolts of the programming have been tested (gyro code, PID feedback, improved camera tracking, holonomic drive), and now that it's working really well, they are now seeing how the design process is supposed to work.