Sometimes I am asked what calculator do I prefer. I like RPN (Reversed Polish Notation) calculators; I find them faster and easier to use than algebraic calculators (after you have accustomed yourself to their twisted way of doing things). This basically limits my preferences to Hewlett-Packard calculators. I have several old models at home, including a functioning 11C, but the two calculators that I use in every day work are the 48GX and the 32sII (an improved version of the 32s).
The 48GX is a very powerful calculator, but I find the 32sII smaller, lighter and easier to use, so most of the time I use exclusively the 32sII. In addition, the display of the 32sII has much better visibility than that of the 48GX. In general I have very little use for graphing calculators, and I tend to use Mathematica or other computer based programs instead of the 48GX for more sophisticated applications.
The 32sII is a programmable calculator, but with very little memory available (384 bytes). Some older models, such as the 42s, are more powerful than the 32sII. I wish Hewlett-Packard would make an improved version of the 42s, call it the 42sII. As things stand now, if you want a non-graphing, chemically pure RPN, scientific programmable calculator, the 32sII is your only choice. Hewlett-Packard stopped making the 42s somewhere in the 90's, before I had an opportunity to get one. Now they are collector items that fetch several hundreds of dollars on e-bay and other trading posts. If Hewlett-Packard ever makes the 42sII, I hope they eliminate the two line display of the 42s and put instead a good visibility display, even if it is only a one line display (as in the 32sII).
The feature of the 42s and 48GX that I miss the most in the 32sII is the capacity to solve linear equations. To partially compensate for this, I wrote two small programs (linear3 and linear4 below) that allows me to use the 32sII for systems of up to 4 equations with 4 unknowns, and still leaves a few bytes free for a small equation or program (such as 2Dvect+, or VectorX below).
