On Mon, Feb 01, 2016 at 12:00:47PM -0800, Danny Yoo wrote: > Here's a response I sent to Srinivas yesterday to further explain why > a balanced binary tree is probably overkill for the "largest > denomination" selection problem. (I didn't realize that I had not > sent the following to the list.) [...] > So there are several crazy avenues we can take to over-optimize this > problem. Just to make it clear: I think sticking to a simple linear > scan makes the most sense. Everything else just seems to try to make > the problem harder than it deserves to be, akin to trying to take the > size of a rectangle via integration. > > http://homepage.usask.ca/~blb230/Math_Comics/Calculus_Comic_files/image001.gif
I'm glad you've forwarded the message to the list, because I love that comic. The clever thing is that it actually gets the maths right too. Some of the notation is a bit strange compared to what I'm used to (I've never seen anyone use a bare integral sign before, with no integrand), and I think he skipped a line, but that's definitely one to keep. Thanks, -- Steve _______________________________________________ Tutor maillist - Tutor@python.org To unsubscribe or change subscription options: https://mail.python.org/mailman/listinfo/tutor