Hacker News new | past | comments | ask | show | jobs | submit login
Street-Fighting Mathematics is published under CC [pdf] (mitpress.mit.edu)
89 points by realitygrill on July 26, 2010 | hide | past | favorite | 26 comments



I think the book is interesting, but I'm slightly disappointed it really wasn't about street-fighting.

I think a lot about mathematics education is a domain issue. Problems are too easily wrapped up in tedious domains that don't engage the imagination. For one of my nationally-assessed Maths projects at school, the teacher had us analyzing football (soccer) scores, looking at standard deviations and the like. The class was far more engaged with the math than I'd ever seen them.

All boys dream of being a bad-ass, so I'm sure mathematics in the domain of street-fighting would also work! ("A perp is able to accelerate his fist at 10 m/s, and his fist has a weight of about 0.5kg. A broken jaw requires 4N of force. How much force will he put in your face? Will he break your jaw?")


Can anyone comment on what would be the 'prereq' for this? Are the video lectures available anywhere (he seems to have taught a class by same name before)? And, approximately how much time would it take for someone going over it as self-study? I know the answer to the last question probably depends, but hopefully a ball mark estimate can be given


It looks like an understanding of Calculus and basic Physics is all that is required to read this book. Ballkpark estimate (how ironic, having to do a ballpark estimation before having read a book on how to do ballpark estimates): between 10 and 30 hours.


When do you people in the US start to learn about integrals?

Is it expected that a first-year undergraduate student already has the knowledge to understand the book?

I don't have the required "prerequisites" as the GP puts it to understand the book, but I should :( and it might shame me into studying a bit :) (then again, I probably won't)


(Single) integrals are freshman calculus, which is nominally taught in the first year of university. In reality, my experience is that many students (and, I would guess most science/engineering/math majors) take AP calculus during their last year of high school (17-18 years old), and proceed to differential equations, multivariable calculus, or whatever is next in their major's sequence during the first year of university.


Thanks for the answers. For all the flak US education receives, the option of taking AP classes sounds good (and it gives college credits, which is even better!).

From your post I could find http://en.wikipedia.org/wiki/Advanced_Placement_Calculus which shows that AP Calculus does indeed include integrals.


Thanks for the answers. For all the flak US education receives, the option of taking AP classes sounds good (and it gives college credits, which is even better!).

This is utterly derailing the main topic of conversation, but...

I know that a lot of mathematicians dislike AP Calculus. The original thinking was that Calculus couldn't be taught to highschoolers, so you waited until University to take it. Aspiring mathematicians would take a rigorous, proof-based Caclulus course, which would prepare them to tackle harder subjects in the future. Everyone else (engineers, chemists, physicists, etc.) would take a more general/applied course. Now, all but a handful of universities offer such courses, under the assumption that anyone who wants to be a mathematician has surely taken AP Caclulus. So the idea of proof-based Calc. for future mathematicians has been lost in transition, and the end result is that you have kids hitting Multivariate Calculus and Differential Equations who haven't seen a proof in their lives.

The current AP Calculus courses are 99% computation. I was challenged while working through Spviak's book with no teacher guidance in highschool, but I scored a perfect 5 on the AP test with little effort.


In short, the lack of a proof-based class renders students Calc-clueless. (groan)

This happened to me. I went straight from high school calc into college differential equations because the AP score allowed me. It took a Rudin-based analysis course, much later, for me to appreciate proofs of convergence or epsilon-delta arguments, because my H.S. calc did not have them, and the college diff-eq assumed you knew them already. The shock was painful.

Eventually though, you learn what you need to know.


It took a Rudin-based analysis course, much later, for me to appreciate proofs of convergence or epsilon-delta arguments, because my H.S. calc did not have them, and the college diff-eq assumed you knew them already. The shock was painful.

I have yet to take a course that uses the so-called "terse little blue book from hell". (:

Eventually though, you learn what you need to know.

Indeed, although I wonder about people becoming discouraged about being mathematicians simply because they've been misled for so long about what's on the "other side" of college math.


you have kids hitting Multivariate Calculus and Differential Equations who haven't seen a proof in their lives.

Really? My AP Calc class was not very proof-focus (probably because the AP Calc test was not), but 10th grade (~15-16 y.o.) Geometry was mostly just proofs. As was Trigonometry. This was 10 years ago, but those same teachers are still at my old school. Presumably they teach the same material. I guess this is atypical? That's too bad, because those trig proofs were actually kinda fun.


I think for most of us the better question is: How soon do you start forgetting all the required details about integrals...


I started doing integrals on my own when I was 14 (grade 9), but my classmates didn't learn about integrals until grade 13. This was in Italy however, and I'm not sure about when they're introduced to students in the States. I can tell you though, that a first year undergraduate student at a scientific faculty shouldn't be overly worried about approaching a book like this.


10 to 30 hours? How did you arrive to that number? The text seems to have many exercise problem. I was expecting an answer like '2 months, going with an hour a day'.

At any rate, thanks.


30 hours means 4 pages an hour. Even solving a few problems here and there, it's hard to imagine you'd proceed at a slower pace than that.

That said, I didn't see that each chapter has a bunch of extra problems at the end. These may take you extra time. I still believe that it's a few weeks project at best, not months.


just read it, most of the material should be accessible to someone with just a decent HS level (up to ~ calculus) math background.


A highly recommended read.

Not to spoil anyone's fun but gems like this abound:

A valid economic argument cannot reach a conclusion that depends on the astronomical phenomenon chosen to measure time.(Discussing GDP and Multinationals in Nigeria).


Looks great! Now can someone get me more hours in the day?


Unless you are a very unique case, you'll should be able squeeze a book this size into your daily life. Let's assume that given the content and the font-size, an average HN reader would require an hour to read 10 pages (that's underestimating most people's reading speed of course). Then the book would require, on average, about 12 hours to read. If you dedicate 2 hours a day to it, you'll have finished it before the week is over. If you take the time to solve all the problems presented, it may take you a few more weeks, but it's not a major project like reading and doing all the exercises from SICP.


This issue isn't so much the time to read the book as the time to read all the books I want to read. I already have two texts underway ("Concepts of Nonparametric Theory" and "Prediction Learning and Games").


I'm very excited :) The book is a combination of developments from the author's phd thesis and a likewise named course that has been run the past few years at MIT's iap term, and is chock full of great tricks and habits to do sophisticated calculations mentally (or at least with greater simplicity)!


Do you have links to those? I know of the MIT ones, but I've also seen links to his Order of Magnitude Physics materials (at Caltech) floating around before.


Why link to the PDF? Why not link to the book's page?

http://mitpress.mit.edu/catalog/item/default.asp?ttype=2&...


I thought it might be a bit hard to notice the pdf link - I'd visited several times but only noticed the CC license because of an Amazon review.


Anyone have a epub version for the iphone/android ?


Since its a small pdf it looks great on my kindle. I believe I read an earlier draft of this and I was impressed.


I would much rather have this book in dead tree form. It seems like it would be great to sit on the couch with.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: