Re: Re: Newbie question: a whine about "professors"
- To: mathgroup at smc.vnet.net
- Subject: [mg9351] Re: [mg9297] Re: Newbie question: a whine about "professors"
- From: Luci Ellis <elisha at dot.net.au>
- Date: Sat, 1 Nov 1997 03:33:37 -0500
- Sender: owner-wri-mathgroup at wolfram.com
Dennis Wayne Scott wrote: > The program iteratively >solves a linear matrix... yeah, I know, "LinearSolve"!!! Tell it to >the professor... > Does anybody else on this group think that teaching one's students to iterate a _linear_ function to find its fixed point is a complete waste of time and an insult to those students? Dennis has Mathematica at his disposal, and so can use LinearSolve, Solve or even the FixedPoint command to get the answer efficiently and with a minimum of fuss. Why, then, does his professor insist on him taking the iteration approach? Maybe the other students don't have access to Mathematica, maybe the professor hasn't heard of it (unlikely in electrical engineering, endemic in economics). But even some other root-finding algorithm would teach the students something a bit more advanced than a blunt-instrument iteration, surely? I know most of the people on this group / list are not economists, so this reference is lost on you <grin> but Kenneth Judd recently published a paper on Computational Methods in Economics where he held an anonymous academic up to ridicule, specifically because said academic claimed that iteration of a particular function was the _only_ way to find its root (in fact, a Newton method would have been much quicker, Judd argued). Perhaps because I am not a mathematician I have missed something here, but are there any cases (other than, perhaps, highly nonlinear and non-analytic functions?) where the only way, or the most desirable way to find the root / fixed point of the function is to iterate it? Your comments are most welcome by email or on the group. Kind regards, Luci Ellis -------------- Luci Ellis: elisha at dot.net.au http://www.dot.net.au/~elisha