It's not unusual that different guesses would get different results, but they should all be within the specified precision. The reason I didn't allow the guess as a parameter is that I didn't want to clutter up the user interface w/ impl. details (I figured that since it's impossible to get exact answers, precision was legitamately a part of the interface, but "guess" wasn't, because there may be a way to write these without the guess. Anyway, I have a feeling there's a more natural way to put together the problem than the way I did it, but I haven't thought of it yet. Jon On Wed, 13 Aug 2003, Philip Ansteth wrote: > > Jon, > > I'll buy your solution. I didn't try that exercise, nor 1.45. > > However, I am curious about something. I get slightly different answers > with different initial guesses. Or I think that's what's going on. Maybe > you can explain it easily. Here's what I did to get the following: > > You have 5 as an initial guess. I thought, Why not pass that as a > user-definable > parameter? Besides, I wanted to make sure that that's what the 5 actually > was--although > later I saw that you had commented to that effect. > > guile> (square-root 33 .001) ; your routine as is > 5.74456265264727 > > Next I changed the argument list: (define square-root (lambda (number > precision my-guess) > And changed 5 to my-guess. That yielded the same result with 5: > > guile> (square-root 33 .001 5) > 5.74456265264727 > > But when I tried other numbers I got slightly different results. > > guile> (square-root 33 .001 1) > 5.74456264846024 > guile> (square-root 33 .001 3) > 5.74456274840839 > guile> (square-root 33 .001 -1) > -5.74456264846024 > guile> > > I didn't pursue it any farther than that. > > I'm trying to keep moving forward. The summer project is about to run out of > summer, > with 500 pages still to go. Today I'm going to look at electrical > resistance problem > in Section 2.1.4, first to see if it looks like a worthwhile project, and > then maybe > to try it. > > Philip > >