Lauren Weinstein <lauren@vortex.com> (05/25/91)
As others have pointed out, extensive Bell Labs research was directed toward the topic of keypad ordering in the late 50's. I distinctly recall the BSTJ article discussing the results. Improving both speed and accuracy were the important factors. An amusing point is that they tested a *very* wide range of configurations. Not only were adding machine style and the current style tested, but a variety of more obscure arrangements. For example, they tested having the buttons in two rows, with both vertical and horizontal orientations tested (the two row configuration was the standard operator MF keypad configuration, by the way). They also tested having the buttons in a circle in the same positions as the finger holes on a dial phone. As I recall this was found to be a pretty "bad" configuration in terms of usage, but today you can find phones with that exact configuration on the market! The issue of the current layout being opposite that of the adding machine was explicitly addressed in the article, I believe. It is important to remember that prior to the advent of the first "four-banger" digital calculators (e.g. the "Bomar Brain" -- remember when they appeared and cost over $100?), adding machines/calculators of any sort were big, mechanical, noisy, expensive, and generally only found in businesses (they also tended to create significant amounts of electrical interference from the motors inside, but that's another story). They were not commonly used by most people, so the adding machine keypad layout was a non-issue as far as the bulk of the population was concerned at that time. --Lauren--