Data processing:Disadvantages of computer data processing

Disadvantages of computer data processing

Computer data processing obviously has many advantages over manual processing. Many offices, in particular small offices, still retain manual methods, sometimes with good reason. For computers are not without their disadvantages:

• They have to be paid for. A small business system, with a decent printer, software, etc., will probably cost around two or three thousand pounds.

• Existing procedures have to be converted to computer processing. The change-over can be protracted, and can cause many headaches.

• A breakdown in the equipment can bring the entire office to a halt.

• Computer processing is more inflexible than manual processing. Once set up, the computer procedures cannot be as easily changed to cope with the changing requirements of a business.

• Computer security can be a major problem. Businesses have been defrauded of large sums of money by employ­ees tampering with their computer systems.

Apart from computer security, these disadvantages are declining in importance with the steady reduction in costs and improvement in reliability and features of computer systems.

Coding systems

We are all familiar with coding systems. Examples include postcodes, employee codes, and National Insurance codes. In a manual DP system it is not essential or necessarily desirable to code data. A clerk may find it easier to deal with a customer or a supplier or a stock part by name rather than by code number. Furthermore, he or she can cope with variants of names, being able to recognize that all refer to the same individual. A computer, however, will not recognize this, and one advantage of using codes in a computer system is that errors that might arise from ambiguities of this sort are avoided.

A second advantage is the small number of digits in a code number compared to the relatively large number in, say, a customer's name and address. Not only do code numbers reduce the keying-in time, they occupy less com­puter storage. This can be important when a data entity, such as a customer or a product, is referenced in several files.

There are three types of coding system, employing either purely numeric characters, or purely alphabetic characters, or a mixture of both:

• Numeric codes offer fewer possibilities for a given length of code (only 10 possible characters per space), but they provide for a check digit that can be used to check the accuracy of data entry (see page 198). Examples of numerical coding systems include the ISBN system used to identify books, and the account numbering system used by banks.

• Alphabetic codes offer more possibilities (26 per space), and they can be descriptive. For example, the alphabetic characters used in air-flight codes indicate the desti­nation (LN = London, NY = New York, and so on).

However, check digits cannot be used with this type of code.

• Alphanumeric codes offer the most possibilities (36 per space), and can include descriptive alphabetic elements.

Accuracy control

Computers don't make mistakes, and one major advantage of computer data processing over manual methods is the accuracy of the output. However, any errors that creep in when data is input to the system will remain undetected. The constant monitoring of the data by humans which occurs at all stages of manual processing is absent. So in an electronic DP system the whole emphasis of accuracy con­ trol is at the data entry stage.

One data entry check that I mentioned above is the check digit that may be appended to a numerical code. This will have a mathematical relationship with the other digits in the number, so that when the number is keyed in the computer checks that the relationship holds. This is an example of data validation, a term meaning the process of checking, by means of a computer program, whether data is valid, i.e. permissible. If, in the check digit test, the final digit of the number does not have the required mathematical relation­ ship with the other digits, then the number is invalid.

Data may be valid, but still wrong. It is conceivable that the data entry clerk may incorrectly key two digits of a code number and yet the required mathematical relationship between the digits holds (in other words, a self-cancelling error has occurred). A data verification check overcomes this possibility- this checks that the data is correct. Unlike data validation, data verification checks can be expensive, for they usually involve keying in the data twice, first by one data entry clerk and then by another.

Because of the importance attached to the accuracy of data, data validation and data verification are often applied at the same time.