Just a two or three years ago (no need to check) I was in college taking computer science courses, and I encountered CS 375 (I think). This was the undergraduate computational mathematics course, and it was taught by a doctoral student.
A large part of the class was about how to calculate (or bound) the errors in computations. It involved things like the condition number of matrices, and I recall thinking the material was so boring!!! In fact, I had a really hard time getting into the class… and got the only bad grade I ever got in a computer science course (but I did pass).
Delightful aside: My revenge was that I received my BS, MS, and graduated with my Ph.D. before the doctoral student who taught the class did. Honestly I blame her advisor(s); she’s an awesome person, really smart, and someone I miss now that she has moved away to fame and fortune. (She’s a college professor, so “fame and fortune” are probably both relative.) I just looked her up and read her course reviews and student ratings. Is it bad that I feel mildly vindicated?
Once I was done with the class I was done with all that condition number, stability, machine epsilon crap! Hooray!
Of course, I subsequently became trapped in the world of statistical testing and reliability analysis, and ended up needing to go back and actually learn that stuff… and then write code to compute it. Once I was done, I promptly forgot it again.
Now that I am rewriting the computations book, I… have to re-learn it once more. Because I’m an idiot. I mean, I could have learned it way back when I took the actual class where I was supposed to learn it… but it was summer, I was in love, and really… it’s still boring.
I am not sure what the moral of this story is, but I can assure you than once I get through re-explaining this in the computations paper… I will undoubtedly forget it once again. I guess this means the moral of the story is that I never learn, but I do occasionally re-learn. When I have to.