TI vs. HP

All, I?m siding with the peacemakers - if you find you think better in stack, use a stack-based machine like an HP. If you find you think better in parentheses, TI/Casio is clearly the way to go. I?m truly glad both exist. I have a couple of observations, though. 1) TI-59 and HP-41 were both powerful enough to program to emulate the other style. I had a relatively easy job programming my TI-59 to run RPN, using the A..E keys for the operations; my high-school friend with the HP-41 had a tougher time programming his machine to do parentheses, using function keys for ( and ). I picked the TI because at the time (~1979), it had nearly the same performance for a considerably lower price. However I now have awful keybounce issues on my -59 and my Dad?s HP-41CV is still going strong. My personal preference both for ease of use and (particularly) for programming is that the RPN machine is easier, but I do concur it takes a while more to get used to. However the difference was less important to me than the cost difference at the time. 2) I quibble with Tony?s recommendation to reject a machine that says Sin(Pi) = 0. I?m pretty sure the TI says that; the quicker but essentially equivalent test I always used to taunt my HP-41-equipped friend was (Sqrt(2))^2. The TI said 2, the HP said 1.99999? I claim both answers are correct. The HP is correct because the rounding error did appear, and the calculator correctly reflected its effect in the final result. However the TI answer is *also* correct because the TI does arithmetic to 13 digits, and displays only the high-order 10 digits. The rounding process from the truncated result to the displayed digits results in the 2.00.. answer which is displayed. I?ll freely admit that the 13-digit-calculation to 10-digit-display rounding process is concealing the truncation problem from me, and if I don?t know to look carefully for it (which can be done, by calculating (Sqrt(2))^2 -2, resulting in 1E-12 or so) I could be bitten badly by it when it finally *does* accumulate up into the displayable digits, or when I do an X=Y test that ?looks? like it should succeed, etc. But, I claim knowing this is part of being familiar with the tools you use, and incumbent on the user. Different tools, different characteristics; both powerful and effective, in my opinion. - Mark