Saturday, October 4, 2008

my Taylor series results

Here's a link for output from my Taylor series program — the detailed version I wrote for my own tests. Once you get the adaptation of the book's code done *correctly* (ouch!), it really does work for sine and cosine of any real x. Without looking at the details for each sum and term, though, not to mention really thinking about it, I would not have found my errors.

I have tested values from all four quadrants of the plane in a variety of magnitudes, although only a few are shown here. You can compare results with the ones a calculator gives you. Note that a TI-89 will calculate a Taylor polynomial of given degree; there are TI-89 emulators online that look and work just like the real thing.

The program that you should submit needs only show the bit that follows "recap" up to "do another": you may wish to write two complete programs, one with all of the detail and a second that cuts the detail after the first one is correct. In that case, submit two .jars, one called Taylor.jar and the other, TaylorDetail.jar. Include the sources in both cases.

No comments: