physical- and psychological hardships for system users.
For example, keyboards that aren't separate from screens can be harmful to the back; screens that don't adjust vertically or horizontally, or that don't have adjustable contrast, can result in increased glare that can lead to a wide range of visual problems. Many products do offer such flexibility, although these aspects are rarely made prominent in manufacturer literature, if they are mentioned at all.
The physical working environment can cause problems, too. For example, the lighting conditions needed for VDT work are far different from those needed for typing on traditional typewriters and doing filing and other typical office work. VDT users need subdued lighting. Moreover, glare from white walls, shiny desk tops and windows often contributes to high levels of glare that make it difficult for employees to work efficiently.
Such considerations may seem less than earthshaking to an executive burdened by price/earning ratios and investment tax credits, but they're issues that you'll have to face in the coming years of the increasingly computerized workplace. But with a little planning, the savvy executive can easily nip such problems in the bud. Perspectives from three users:
PPG Industries "As the result of an internal study of office technology, we installed a minicomputer system, with fourteen work stations," says Ken Myers, supervisor of programming support for the chemical division of PPG Industries in Pittsburgh.
"During the first few months, we're concentrating on the people aspect of the system, making sure that the capabilities are being fully utilized, that the training is sufficient and that the personnel is adequate. We've started the system primarily on word processing, although we intend to explore some of the additional advantages as we get moving.
"In our original investigation, during which we talked with a number of computer companies, the discussions always seemed to deal more with people than with technology. We realize that it was important to pay due respect to the human element when going into something as new as word processing. We kept close touch with the personnel aspects of our system. We had orientation sessions prior to the actual installation. We made sure that everyone had training that was offered through our vendor, so that when the machine showed up they had some background on what it could do and had the ability to start using it.
"So we tried a combination of all those things to take a little pressure off the secretaries, and at the same time see them accelerate and get through the learning curve as soon as possible, without expecting full proficiency for six to eight weeks. I think that it has been a big success, because our people have almost unanimously indicated that they really enjoy working with the equipment. They feel that it has been a big benefit to them."
United Mineralogies Inc. and
"Our core group of companies is involved in many activities, including active real estate development, agriculture, and the development of oil, gas and coal properties," says Daniel F. Flynn, president of United Mineralogies Inc. and Jonnycake Inc., two jointly owned companies based in Farmington, Connecticut. "Our interest in data processing was instigated by the need to assemble raw data from our various operations around the country, and to be able to analyze these operations and the opportunities that they present in order to determine which are most deserving of our attention. Our data-processing operation, therefore, is broken down into two parts: the gathering and assembly of data, whether statistics or financial reports, and the manipulation of this data into usable form.
"I would recommend highly that any executive who is involved with analysis spend a couple of days taking an executive computing course. Not simply to become a proficient data-processing operator, but to understand the dimensions and capabilities of the current state of the data-processing art.
"We purchased terminal equipment but decided to go with a service bureau for the software because the technology is changing so quickly that by the time we invested or committed ourselves to one system, developments would have taken place that would leave our equipment out of the top of the market. So we chose for flexibility to go for terminal equipment that is plugged into the service bureau's larger system.
"Our uses and benefits have expanded almost geometrically over the past couple of years, and we couldn't be happier. I suppose that there could be a point in time when we might purchase our own system, but for the volume of information that we store, it simply isn't cost effective."
"About a year ago we realized there were numerous cases where we couldn't obtain routine data for reports," says E. C. Yerian, manager for industrial engineering and distribution planning for International Harvester in Broadview, Illinois. "We knew that the data was available, but we'd go to our data processing people and have to write out a project request and go through a lot of bureaucratic red tape. It was taking us weeks to get what we needed. The DP department was concerned with maintaining inventory of the five million square feet of warehouse space we have scattered around the world, so we got the idea we could do it ourselves.
"We went out and purchased a small microprocessor and started learning basic programming. All of a sudden we had about six departments that wanted to use it. Within a year, the thing was in use from 6 A.M. to 10 P.M.
"After one year, we expanded our processing capacity. We purchased a system with five terminals. It's totally operated by the users-representatives from the supply department, the accounting department, the transportation department, engineering.
I think that we have eighteen steady users on it now, and that number is growing by the hour. The data processing department, which is still running its big mainframe system, has offered to give us support, but so far we haven't felt the need for any assistance.
"It's not a duplication of effort. The functions that we are running on this system are entirely our own creations, like simulations and employee reports that we used to type manually. Now, once we key in information, we can quickly manipulate the data and restructure reports for any level of management."®
For Models I, II, III; tape, disk or stringy floppy
Last year we compared two BASIC compilers for the TRS-80 Model II: Microsoft's BASIC Compiler (BASCOM) and Radio Shack's Compiler BASIC (RSBASIC). They are now available for the Model I. In addition, a third compiler, ACCEL2 from Southern Software in England (via Allen Gelder Software in this country), has made its appearance.
The usual reason for using a compiler is to gain speed. This is not always the case. CBASIC, long a favorite of CP/M users on non-Radio Shack computers, is not fast. Even on machines with identical processors (Zilog's Z-80), the built-in BASIC in the Model I can outrun CBASIC in speed tests.
Another reason for using some compilers is for accuracy. Put the following code into your computer and run it. 10 A = 0.1 20 B = B + A 30 PRINT B 40 GOTO 20 Now you have a stream of numbers running down the video. Everything is accurate to the nearest tenth of a whole number, right? No?
Now make A = 0.01 in line 10 and run the test again. Same results? No, they are different this time. At any rate, it doesn't take long to see that accuracy is not guaranteed with standard Radio Shack BASIC (which we will simply call Standard BASIC for the remainder of this article). Many versions of BASIC are plagued with the same problem. Some decimal numbers cannot be accurately represented as a binary number. In most cases the conversion process from decimal to binary and back takes care of the inaccuracies. However, when two binary numbers are added, the inaccuracy can become greater than the binary to decimal conversion can handle. Hence, the eventual loss of visible accuracy.
In business applications where numbers must balance to the penny, this inaccuracy eventually becomes intolerable. Two possible solutions present themselves: Use a language which works in binary coded decimal, or run all numbers through a self-correcting subroutine. Usually the former solution means COBOL and the latter means number to string and back again, with a loss of time.
The following routines demonstrate a double conversion correction routine. Note that this routine does not always work. Try the double precision routine with A = 0.0001. The first listing is for single precision numbers and the second is for double precision.
Was this article helpful?