Data entry has traditionally been the weakest link in the chain of data-processing operations. Although it can be processed electronically at extremely high speeds, significantly more time is required to prepare the data, enter it, and then check its accuracy before it can actually be processed. Another method of data entry, source-data automation, collects data about an event, in computer-readable form, when and where the event takes place. This eliminates the possibility of keying incorrect information into the computer. Source-data automation improves the speed, accuracy, and efficiency of data-processing operations.
Source-data automation is implemented in several ways. Each requires special machines for reading data and converting it into machine language. The most common forms of source-data automation are discussed in the following paragraphs.
Magnetic Ink Character Recognition (MICR)
Magnetic-ink was introduced in the late 1950s to speed check processing in the banking industry. Because magnetic-ink characters can be read by both humans and machines, no special data-conversion step is needed. Magnetic-ink characters are formed with magnetized particles of iron oxide. Each character is composed of certain sections of a seventy-section matrix. The characters can be read and interpreted by a magnetic-ink character reader. This process is called magnetic-ink character recognition (MICR).
With MICR each character area is examined to determine the shape of the character represented. The presence of a magnetic field in a section of the area represents a 1 bit; the absence of a magnetic field represents a 0 bit. Each magnetic-ink character is composed of a unique combination of 0 bits and 1 bits. When all sections in a character area are combined and translated into binary notation in this manner, the character represented can be determined. MICR devices automatically check each character read to ensure accuracy.
Processing bank checks is a major application of magnetic-ink character recognition. The magnetic ink characters are printed along the bottom of the check (see picture below).
The transit field is preprinted on the check. It includes the bank number, which is an aid in routing the check through the Federal Reserve System. The customer's account number appears in an "on-us" field. A clerk manually inserts the amount of the check in the amount field after the check has been used and received at a bank.
All magnetic-ink characters on checks are formed with the standard fourteen-character set shown below.
Other character sets may be used in other applications. As the checks are fed into the MICR device, it reads them and sorts them by bank number at a Federal Reserve Bank and by account number at the issuing bank. In this manner, checks are routed back to each issuing bank and then back to its customers. A MICR device can read and sort hundreds of checks per minute.
Optical Mark Recognition (OMR)
In another form of source-data automation, optical recognition devices read marks or symbols coded on paper documents and convert them into electrical pulses. The pulses can then be transmitted directly to the CPU or stored on magnetic tape for input at a later time. The simplest approach to optical recognition is known as optical-mark recognition (OMR), or mark-sensing. This approach is often used for machine scoring of multiple-choice examinations (see picture below).
A heavy lead pencil is used to mark the location of each desired answer. The marks on an OMR document are sensed by an optical-mark page reader as the document passes under a light source. The presence of marks in specific locations is indicated by light reflected at those locations. As the document is read, the optical-mark data is automatically translated into machine language. When the optical-mark page reader is directly connected to the computer, thousands of forms of the same type can be read and processed in an extremely short period of time compared to entering the data by hand.
Optical-mark recognition is also used in order writing, inventory control, surveys, and questionnaires, and payroll applications. Since optical-mark data is initially recorded by people, forms that are easy for them to understand and complete must be devised. Instructions, with examples, are generally provided to aid those who must use the forms. Good design helps prevent errors and lessens the amount of time required to complete forms.
Bar Code Reader
Another type of optical reader, the bar-code reader, can read special line or bar codes. Bar codes are patterns of optical marks that represent information about the object on which the code appears. They are suitable for many applications, including point-of-sale (POS) Systems, credit card verification, and freight identification to facilitate warehouse operations (see picture below).
Data is represented in bar code by the widths of the bars and the distances between them. Probably the most familiar bar code is the Universal Product Code (UPC) found on most grocery items. This code consists of pairs of vertical bars, which identify both the manufacturer and the item, but not the item's price. The code for each product is a unique combination of these vertical bars. The UPC symbol is read by a handheld wand reader or by a fixed scanner linked to a cash register-like device. The computer system identifies the product, its brand name, and other pertinent information and uses this data to find the item's price. It then prints out both name and price. The computer keeps track of each item sold and thus helps the store manager to maintain current inventory status.
Optical Character Recognition (OCR)
Optical-character readers can read special types of characters known as optical characters. Some optical-character recognition (OCR) devices can read the characters of several type fonts, including both uppercase and lowercase letters. The most common font used in OCR is shown below.
A major difference between optical-character recognition and optical-mark recognition is that optical-character data is represented by the shapes of characters rather than by the positions of marks. However, both OCR and OMR devices rely on reflected light to translate written data into machine-readable form.
Acceptable OCR input can be produced by computer printers, adding machines, cash registers, accounting machines, and typewriters. Data can be fed into the reader a continuous form such as cash register tape or on cut forms such as phone or utility bills. When individual cut forms are used, the reader can usually sort the forms well.
The most advanced optical-character readers can even read handwritten characters. Handwritten characters must be neat and clear, however, or they may not be read correctly. The system is not foolproof because handwriting can vary so much from person to person. The optical-character readers reject any characters that cannot be interpreted. Devices that must read handwriting are often very slow.
Machine-produced optical-character recognition has been used in credit card billing, utility billing, and inventory-control applications. Handwritten optical-character recognition has been used widely in mail sorting. The reliability of optical-character recognition systems is generally very good.
Remote Terminals
Remote terminals collect data at its source and transmit the data to a central computer for processing. Generally, data is transmitted over telecommunications equipment. The many types of remote terminals available can increase the versatility and expand the applications of the computer.
Remote terminals that perform the functions of a cash register and also capture sales data are referred to as point-of-sale (POS) terminals. Such terminals have a keyboard for data entry, a panel to display the price, a cash drawer, and a printer that provides a cash receipt. A P0S terminal typical of those found in many retail stores is pictured below.
Some POS terminals have wand readers that can read either the Universal Product Code (UPC) or the OCR characters stamped on or attached to an item. The sale is registered automatically as the checkout person passes the wand reader over the code; there is no need to enter the price via a keyboard unless the wand malfunctions. Thus, P0S terminals enable sales data to be collected at its source. If the terminals are directly connected to a large central computer, useful inventory and sales information can be provided to the retailer almost instantaneously.
Touch-tone devices are remote terminals used together with ordinary telephone lines to transfer data from remote locations to a central computer. The data is entered via a special keyboard on the terminal. Generally, slight modifications need to be made to the telephone connection to allow data to be transferred over the line (see picture below).
There are several types of touch-tone devices. One that reads a magnetic strip on the back of plastic cards is often used to verify credit card transactions. Another stores large amounts of data on a magnetic belt similar to a magnetic tape before transmitting it. This type of terminal is best suited for large-volume processing.
Last Updated Jan.6/99