Of the two activities, printing and reading, the first to be performed is the printing. After all, if something is be read, it must first be available in a readable form. Someone else may be printing the bar code symbol for you, as in the case of the UPC codes printed on items in the grocery market. Or you may be printing the symbol for someone else to read, such as would be the case if you were providing something to be resold in the grocery market. 

The initial step is to carefully read the specification, especially before the printing equipment is selected and purchased. If no specification exists, then the first order of business is to write one. No matter how simple or controlled your application, you are asking for disaster without a specification. It establishes the ground rules under which everything must operate. Even though your scanner can read the labels your present printer is producing, there is no guarantee this will be the case if you should ever replace either with different equipment, or even if you change ribbon manufacturer.

 Once the operating conditions are determined, it is time to select a suitable method for generating bar code symbols. To do this, it is necessary to understand the limitations of the various technology choices. These limitations must then be carefully compared to what the label is required to do.

However, to realize the benefits offered by bar codes, their capabilities and limitations must be understood. Bar code systems are not designed to duplicate the human visual capabilities, like OCR (Optical Character Recognition) systems, but to replace it with components designed with readily available and low cost technology.

 Because the symbol is being read by a scanner, its quality level should be judged against what the scanner expects to see. Here is an example of "what you see is not necessarily what you get". The human eye may think the contrast between the bars and spaces is excellent, but if they are printed with red ink on a white background they will appear indistinguishable to a laser scanner operating in the infrared spectrum. If the bars are created by overprinting with a heavily inked ribbon on a dot matrix printer, then edge bleeding of the printed bars can result in wide spaces being interpreted by the scanner as narrow spaces. 

 Being able to read a bar code symbol with a scanner is not an acceptable method of determining if it is within the allowable tolerances. A very poor quality symbol may possibly be read with a high performance scanner, but in this case the scanner is allowing the symbol to infringe upon the tolerances reserved for the scanning system. Another scanner, or even the same scanner with a different operator, might not be able to compensate for this lack of symbol quality, rendering symbol unusable. For this reason, the standards will spell out the minimum acceptable levels of contrast, reflectance and other critical print quality measurements. They also specify how these measurements are to be made. 

Verifiers should be used with a heavy dose of common sense. Most quality printers designed for bar coding applications have few problems printing consistent symbols if they are maintained within the proper operating parameters (a good ribbon on a dot matrix printer for example). Therefore a good sampling program can be used to maintain acceptable an quality level. It makes little sense to "verify" each label as it is printed, unless your printer cannot give consistent results. If that is the case, it is time to get a better printer. What is more important is whether or not the symbol can be read in the days or weeks ahead, after it has been sitting on someone's receiving dock for a couple of weeks. 

 The next consideration is for the base label material, which must be matched to both the print technology and the environment. If a symbol is generated by transferring ink to a label surface, then the ink must not only adhere at the time of printing, but throughout the useful life of the label. It must also provide a method of attachment to the item to be identified if it is not printed directly on the component. This is commonly done using an adhesive backed label material, but other methods may be used such as attachment using string or even stapled to the item. Whatever method is used, it is critical that the label stay attached to the item. A label laying on the warehouse floor is as useless as one that cannot be read but still attached to the item. 

 Item Identification - An identification label is generally very simple with only a single printed symbol. The size of the label can be severely constrained by the size of the item, but otherwise, the labels can be sized for the most economical production. The system stores all the data that is identified with the particular code symbol and matches it to the item when the number is read into the system. This is generally referred to as a "license plate" system since it operates in much the same way as the license plate on your automobile. By itself, the license number means very little, but, combined with information in the Department of Motor Vehicles data base, a good bit of information can be retrieved. The UPC symbol is the most common example of this type of system.

 Multi Code Shipping - If the label must contain several different fields, then a larger label is dictated. A typical shipping label is an example of this type. It must contain all of the information required by the sender (invoice number), carrier (shipping address) and the receiver (purchase order number). Each package may contain something dissimilar or go to a different customer, so the label is unique. This requires a printer capable of producing a large multi tiered label that may be read by different types of scanners. One of the problems associated with multi symbol labels is how to know in what order the symbols should be scanned so that the system knows which data goes with which symbol. This is solved by the use of an identification character(s) added to the encoded data. These characters are called "Data Identifiers". The Federation of Automated Coding Technology (FACT) maintains a standard which lists the industry accepted Data Identifier characters. If this standard is observed, the type of data encoded in the symbol, such as a part number, can be determined even if it is being read by another company in a completely different industry. 

 Distance - How far away is the scanned symbol? The particular scanner used will determine the maximum distance and the minimum width of the bars for that distance. Long distance scanners must have powerful light sources and the bar spaces must reflect enough light to be detected by the scanner. A scanner that is very close to the symbol, such as a contact wand, does not have to illuminate it with as much light to get a usable reflection, since the distance involved is much smaller.

 Speed - The symbol must be presented to the scanner long enough to be decoded. Speed variations may range from constant when using conveyor belts to wide when scanning with human hand movements. In general, the faster the speed or greater the variation, the larger the symbol must be since the scanner must be able detect which bars are wide and which are narrow and then decode the pattern.

 Symbol Orientation - Bar code symbols are single dimensional. The information contained can be read only by scanning in an axis along the length of the symbol where the scanning beam can cross all of the elements. It is not like forgetting to dot an "i" or cross a "t", where most humans can still recognize the character. If the scanner does not get all of the information because it misses a bar, it doesn't guess, it just gives a "no read". Bar code symbols are also vertically redundant codes with the information contained being the same if it is scanned at the top or the bottom. If the orientation of the symbol cannot be controlled as it passes the scanning station, there are two choices for handling the problem. Either two identical symbols printed orthogonally to each other must be used, or the scanner must be capable of scanning in either orientation.

 Light Source - Since bar code symbols depend upon differences in reflected light as information, the scanner must have a light source available. It does not have to be a light source visible to the human eye, just one that can be recognized by the scanner. Of the most economical and reliable light sources available, laser diodes are probably the most widely used. Some of these emit infrared light that is invisible to the eye and require a bar to be printed with an ink that will reflect the infrared light. The human eye is incapable of judging the reflectance quality of these symbols as seen by the scanner. Therefore, the reflectivity characteristics of the printed symbol must be chosen to match the light source of the scanner used. 

 How Long - When a bar code label is used to identify an item, its life is not necessarily that of the item. A cardboard box has a shorter lifetime that the TV set inside, yet both are probably marked with the same bar coded identification number. How and for what purpose the label will be used is more important in determining the required life than what it is identifying. Shipping labels are good examples of limited lifetime labels. They may go on air express packages or truck carrier shipments, with attendant usage lives varying from one day to one month. Unless you are shipping to a destination an infinite distance away or using an infinitely slow carrier, the label only has to last a limited amount of time.

 Under What Conditions - No label is permanent; it just depends upon what lengths you want to go to before it destructs. Don't confuse a controlled environment with one that is not harsh. The flow solder machine used earlier is an example of a harsh environment that is also very controlled. The label must remain usable after being subjected to temperatures up to 600° F. Don't forget that after the high temperature, the label will probably undergo some extreme chemical environments at the degreaser station that follows. 

Batch Printing
Batch printing implies that the data to be printed on the label is known far enough in advance to have them printed remote to the using location. Sometimes this leads to increased control problems. If the labels are serialized, for example, a voided or lost label can have serious consequences. On the other hand, if the label is produced at the point of usage at the time it is needed, the chances of it getting lost or placed on the wrong article are greatly reduced. Batch printing can be done in two basic ways, depending upon how much time the label data is known in advance.

Off Site Printing - If the information required to print the labels is known weeks or days in advance and the quantities used are sufficiently large, then it becomes an exercise for the purchasing agent. There is tremendous latitude in label size, materials and supply format (i.e. rolls, sheets, individual, etc.). This is probably the most cost effective method of generating high quality labels, and most likely the only way of getting some specialty labels. However, an inventory level of labels is required to meet the usage requirements. If a large number of different labels are required, each must be inventoried with the accompanying increased likelihood that some will be scrapped as products undergo change.

On Site Printing - This is similar to the off site classification except the exact data for the labels is not known until hours in advance. Typical of this category would be date coded part number labels for a multi product production line, only a few hours notice may be available when the line is converted to a different product. Depending upon the number of labels needed, the choices could range from an in house press to a high speed label printer. There is less latitude in material selection because prep and print time is limited, but a wide variety of printers can be used. The incidence of scrapage is reduced since a smaller number of each label must be inventoried. 

Demand Printing -Printing labels only when they are required, or demand printing, has many advantages. It is also more restrictive on how the labels are produced. Demand printers can operate in the real time production environment. They accept data from the "system" and produce a label with data unique to that particular article. The variable data can be almost anything; serial numbers, sequence codes, test results, date codes, lot codes, etc. Since the labels are used as they are produced, there is no scrappage problems associated with inventory. Because the labels are printed on the spot and usually one at a time, the type of printer needed has to meet certain requirements.

Speed - The label must be ready when it is needed. This can be very fast if an automatic applicator is used, to reasonably slow if a human operator takes the label and places it on the article. At the same time, the print process used must be simple and fast. Multi step printers do not make good demand printers because of the time needed for extra steps and set up.

Presentation - The label must be presented to the "applicator" ready to be used immediately after it is printed. This could mean stripped with the adhesive back exposed or with the backing liner still attached. The last printed label must be easily accessible and cannot be trapped by the mechanism.

Media - The media must be universally usable for all label requirements supported by that printer. While it is possible to change label supplies to get different label sizes, pre print information, etc., it is not feasible to change the media except on supply type basis (i.e. one roll, one sheet, etc.).