1 Mar
2009
1 Mar
'09
2:53 a.m.
rwg:
http://www.tweedledum.com/rwg/pecu.JPG is a variation on one we did here a few yrs back. Illustration by a 13 yr old who then solved it in the most calculus-intensive way possible.
gjm:
What's the *least* calculus-intensive way possible? (Or at least the easiest way possible.) I did it by saying d(area)/d(normal) = boundary_length, as it were (the derivatives being easier to calculate than the things they're derivatives of); the resulting answer doesn't seem quite simple enough for there to be a substantially more elegant solution, but perhaps either my answer or my intuition is wrong.
It turns out that this solution is the one rwg had in mind. -- g