<p>fignewton, I think the calculator for #3 is mostly a speed factor, not for additional understanding or ability. I suppose it could be useful for functions that you somehow knew how to put into the calculator but couldn’t find a derivative for, though I’m blanking on any useful ones off-hand.</p>
<p>baseliner, the College Board website has a list of calculators that are both allowed and calculators that have all the required features ( [AP</a> Central - Welcome to AP Central](<a href=“AP Calculus: Use of Graphing Calculators – AP Central | College Board”>AP Calculus: Use of Graphing Calculators – AP Central | College Board) ), and your calculator is on the list with an asterisk, meaning that it is both allowed and has the necessary features for the AP Calculus exam.</p>
<p>I don’t know how those features work on that particular brand of calculator, and I usually recommend that if you’re going to have a calculator, it should be of the same series that the teacher has, unless you know about all the features on your calculator or can pick them up quickly.</p>
<p>UnleashedFury, I was at a workshop this summer and received a warning about using the graph screen. I can’t remember whether it was for derivatives, integrals, or both, but it had to do with the level of tolerance being used for the functions. I tried looking it up on Google to see if I could find which one it was, and while I found the warning for integrals, I didn’t find it for derivatives.</p>
<p>The reason has to do with the way the TI calculates both derivatives and integrals. The TI-83/84 model doesn’t actually calculate the derivative or anti-derivative, it simply applies numerical approximation methods that are usually equally acceptable. For instance, for the derivative it calculates [f(x+0.001)-f(x-0.001)]/0.002 on the nDeriv menu (although you can change the 0.001 by applying a different value after the x = point). The fnInt feature uses Riemann sums with rectangles 100 times thinner than the integral of f(x) feature on the graph screen. What I don’t know is the tolerance for dy/dx, and if it is different, it could potentially impact accuracy at the required number of decimal places.</p>
<p>(By the way, as a fun little side effect of how calculators work, try finding nDeriv(1/x,x,0). Symmetric quotients are fun when applied blindly.) :)</p>