Chapter 20

Richard E. Monnier / Thomas E. Osborne / David S. Cochran

Many of the day-to-day computing problems faced by scientists and engineers require complex calculations but involve only a moderate amount of data. Therefore, a machine that is more than a calculator in capability but less than a computer in cost has a great deal to offer. At the same time it must be easy to operate and program so that a minimum amount of effort is required in the solution of typical problems. Reasonable speed is necessary so that the response to individual operations seems nearly instantaneous.

The HP Model 9100A Calculator, Fig. 1, was developed to fill this gap between desk calculators and computers. Easy interaction between the machine and user was one of the most important design considerations during its development and was the prime guide in making many design decisions.

One of the first and most basic problems to be resolved concerned the type of output to be used. Most people want a printed record, but printers are generally slow and noisy. Whatever method is used, if only one register is displayed, it is difficult to follow what is happening during a sequence of calculations where numbers are moved from one register to another. It was therefore decided that a cathode-ray tube displaying the contents of three registers would provide the greatest flexibility and would allow the user to follow problem solutions easily. The ideal situation is to have both a CRT showing more than one register, and a printer which can be attached as an accessory.

Figure 2 is a typical display showing three numbers. The X register displays numbers as they are entered from the keyboard one digit at a time and is called the keyboard register. The Y register is called the accumulator since the results of arithmetic operations on two numbers, one in X and one in Y, appear in the Y register. The Z register is a particularly convenient register to use for temporary storage.

One of the most important features of the Model 9100A is the tremendous range of numbers it can handle without special attention by the operator. It is not necessary to worry about where to place the decimal point to obtain the desired accuracy or to avoid register overflow. This flexibility is obtained because all numbers are stored in 'floating point' and all operations performed using 'floating point arithmetic.' A floating point number is expressed with the decimal point following the first digit and an exponent representing the number of places the decimal point should be moved-to the right if the exponent is positive, or to the left if the exponent is negative.