Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

The rigorous version would be:

0 < (Vol(Bε) - Vol(Aε))/ε² = ((a+b+c) - (a'+b'+c'))π + 2((ab + ac + bc) - (a'b' + a'c' + b'c'))/ε + (abc - a'b'c')/ε²

The limit for ε to infinity must be >= 0 and is ((a+b+c) - (a'+b'+c'))π, therefore a+b+c >= a'+b'+c'.

"large epsilon" is a standard way to express "consider the asymptotic behavior for epsilon to infinity and you will see what I mean".



Thanks, I get that it means asymptotic behavior - that's the thing I don't get the motivation for.


I think the intuition is that we want to focus on the edges of the box (where the lengths sum together like the answer we're looking for), not the volume (where the lengths multiply together and cause problems). So we expand the edges until they completely dominate the equation.




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

Search: