Barcodes

Barcodes are everywhere around us, from the back of your cereal box to a part in the latest electric vehicle. They're on the back of books, and the bottom of rocket engines. Anytime you need a machine-readable format to represent data visually.

Here I just want to talk about the ones that I've run into the most, and a bit about them, what they look like, and some details. There's literally hundreds of barcode schemes out there, but many are very narrowly scoped.

Formats

Formats can be broken down into 1D and 2D formats. As the name implies, 1D barcodes require traversing the barcode in a single dimension. 2D barcodes, like QR codes, require traversing the code in 2 dimensions. This also impacts the complexity of the reader required.

1D Barcodes

UPC A

UPC-A is one of the most common and well-known barcode types. It's the one you've seen on the back of a cereal box. In fact, the first UPC barcode ever scanned was a pack of Wrigley's Juicy Fruit chewing gum. It is widely applied to retail in the United States. It is also known as Universal Product Code version A. You can't just generate your own UPC-A code, as the assignment of manufacturer ID numbers is controlled in the USA by the Uniform Code Council (UCC), which is now just known as GS1 US.

UPC-A is a subset of EAN-13 (the European counterpart and largely interchangeable). To be specific, an UPC-A bar code is an EAN-13 bar code with the first EAN-13 number system digit set to "0". Beyond this, UPC-A barcode is able to convert to UPC-E. A UPC-A barcode consists of 12 digits. It begins with a single digit number system character, which designates how the code should be classified: as a regular product, a weighted item, pharmaceuticals, coupons, etc. After that is a five digit manufacturer's number, followed by a five digit product number, and finally a check digit. Each digit is represented by a uniquely identifiable pattern of two bars and two spaces of varying width. No letters or other special characters aside from numbers may be used.

A mathematical formula computes the check digit from the other numbers.. So if the check digit doesn't match the digit generated by the other numbers, you know immediately that there's an error.

Advantages

UPC-A barcodes are simple, short, popular, and can be read by just about any standard barcode reader. In addition, the check digit guards against accidental errors, especially when entering in the code by hand.

Disadvantages

The 12 digit system of a UPC-A barcode provides for small, limited encoding, which makes it ideal for supermarket and retail use, but less suited for other scanning and identification purposes. You can't just reuse it and let your own codes out in the wild without having assigned numbers.

Code 39

Code 39 barcodes are also called a Code 3 of 9 barcode. I believe it was the first barcode to allow the use of both letters and numbers. It's a variable length barcode that can have up to 43 alphanumeric characters.

A Code 39 barcode contains both a start and end symbol for the scanner. Typically, the scanner will convert this to a *. Aside from the start/end character, the barcode can technically only encode the numbers 1-10. To encode letters, it uses special designations. It does this by dividing them up into categories:

• For example, the first 10 letters (A-J) are given numeric values, preceded by a "Letters" designation.
• The next 10 (K-T) are designated "Letters +10." So K would be Letters +10, followed by a 1, showing that it's the 11th letter of the alphabet.
• The last set (U-Z) then become "Letters +20."

Numbers, of course, have their own designation as well.

Advantages

The use of both letters and numbers makes Code 39 more versatile. In addition, it is self-checking, so it does not require a check number (though one is still recommended).

Disadvantages

It is limited to a maximum of 43 characters. Also, its method of assigning number values to letters in order to read them limits versatility and precludes other characters. It also makes decoding the barcode more complicated.

There is a related Code 93 barcode that is used by Canada Post that has more density, but is otherwise very similar.

Code 128

Code 128 is a denser barcode than most, but it is also the most versatile barcode in the 1-D category, in terms of potential information storage. Its name comes from the fact that it can encode all 128 ASCII characters. That includes letters and numbers, but also punctuation, symbols, and more. It is most commonly used in the logistics of things like purchasing and shipping, but can potentially be used for a variety of other purposes.

There are six sections to a Code 128 barcode. The first and last sections are both quiet zones, which consist of a certain amount of white space, based on other elements of the code. After the opening quiet zone is a start character, which designates what code set the barcode falls into. Code 128 has three code sets:

• Set A designates codes with all capital letters.
• Set B is for codes with both capital and lower case letters.
• Set C is for codes with only numeric data -- which allows it to compress twice as much data into the same space.

After the start character comes the data itself. Each encoded character consists of exactly three bars and three spaces. After that is a check digit, to ensure accuracy, followed by an end character to signify the end of the code, and finally the closing quiet zone.

Advantages

Not only is Code 128 versatile, it is compact, being able to store a much larger amount of data than a regular barcode in the same amount of space.

Disadvantages

Mostly just the limitation to the length of the barcode itself and therefore he length to the amount of data you can encode.

2D Barcodes

PDF417

PDF417 is used in a variety of applications, primarily transport, identification cards, and inventory management. It is defined in ISO 15438. PDF stands for Portable Data File and was developed by Symbol Technologies. As far as I'm aware, it has absolutely no relation to Adobe Acrobat's PDF, or Portable Document Format. PDF417 uses built-in error correction to ensure better readability, and it can encode 100-200 characters quite comfortably.

PDF417 is a different style than QR codes and Data Matrix in that it is not strictly a 2D spatial orientation, but instead a series of stacked 1D barcodes. A single PDF417 symbol can theoretically hold up to 1850 alphanumeric characters, 2710 digits or 1108 bytes. Due to internal data compression algorithms the exact data capacity depends on the structure of the data to be encoded.

The barcode contains a way to detect and correct errors based on the Reed Solomon algorithm. The level of correction included can be between 0 and 8.

To give you an idea of the age, PDF417 uses the CP437 character set that was introduced with the original IBM PC in 1981. It also uses a very odd codewords approach to encoding.

Advantages

As far as I can tell, the only advantage of the PDF417 barcode is that it can be read with a "simple" line reader that's swept over the barcode. That was likely a huge benefit in the original timeperiod, but now is simply not worth the hassle.

Disadvantages

Complexity. Ooof. It's also about 1/4 the data density (bits per square centimeter, for example) of either a QR code or a Data Matrix.

QR Code

QR (Quick Read) code is a matrix code originally designed for use in the labeling and identifying of automotive parts. It's became very popular with the introduction of camera phones that could easily read it.

A QR code can be either very simple or very complex, and can vary in size. It's a square shape that includes both black and white cells. The top two corners and the bottom left corner each contain a small finder pattern, displayed as a square within a square.

QR codes can hold a maximum of 2,953 bytes, 7.089 numbers, or 4,296 alphanumeric characters. There is also a "micro" version of the QR code that can hold anywhere from 5-35 numbers, 5-21 alphanumeric characters, or 7-15 bytes.

Advantages

It can be read very quickly and has an enormous storage capacity, making it superior to UPC bar codes in just about every way. It can encode both numeric and alphanumeric characters, as well as binary characters and even Chinese logographic characters.

Disadvantages

It has a high storage capacity, as well as error correction for damaged codes, but it's a tradeoff between the two. The higher the error correction, the lower the storage capacity. Also, since the standard allows it to specify that it contains executable code (Javascript via URL), there are some additional risks that simply aren't required.

Data Matrix

The Data Matrix is a 2D matrix code (ISO/IEC 16022), capable of encoding very large amounts of data (2,335 alphanumeric or up to 3,116 numbers)in a compact space. Data Matrix codes are made up of small black and white squares that form a big square or rectangle. They're used in a variety of industries, including aerospace, component labeling, food and beverage, pharmaceutical, defense, mail, and printed media where tracking and traceability is critical. It is also a core part of the GS1 framework for global trade.

These are my favorite barcodes (not really a barcode, but...), if one can have such a thing. They're compact, reliable, and don't waste a ton of space on things that others do.

Each Data Matrix has multiple parts:

1. Finder or 'L' pattern. This helps a barcode reader locate and determine the orientation of the code.
2. Clocking pattern. This provides a count of the number of rows and columns in the code.
3. Data region. This can be alphanumeric data. Redundant data using Reed-Solomon encoding, which can make up anywhere from 28% to 62.5% of the available data, is often included so even if one or more cells are damaged, the code is still readable.
4. Cell or module. This is the blocks inside the data region and are the interior components of the code that make up the pattern
5. Quiet zone. The margin of white space around the entire code; Data Matrix codes must have a quiet zone to be decoded correctly a perimeter finder and a timing pattern.

While there are a lot of different Data Matrix arrangements, let's talk about just a few of them: 10x10, 32x32, and 144x144 (the largest). Each of these is a different number of "regions", which are the maximum size before you have to reset the finder pattern. We'll look at how much they can hold, and how that is broken down by data and error recovery.

There are a huge number of encoding options for a Data Matrix, including:

• ASCII (ISO/IEC 646)
• C40, which is optimized for uppercase letters and numbers. There's some information here.
• X12 EDI for heavily structured and hierarchical data. Mostly, I've run into it in the healthcare space.
• EDIFACT, a United Nations standard for EDI, so you know it's going to be simple and uncomplicated. It is, however, very widely used in B2B interactions.

It's worth noting that the standard use of Data Matrix is based on the GS1 Data Matrix guidelines, which institute some additional structure to it. One of the neat things about Data Matrix is that it is designed to be applied in a wide variety of technologies, including direct inkjet printing, laser engraving, and even dot peen marking,

Advantages

The barcode is designed to be read even when it's up to 30% damaged, due to a built in error correction system. It's also capable of encoding either letters, numerical data, or other ASCII characters. Data Matrix codes can be read with image-based barcode readers or mobile devices. It can operate with a lower resolution, and is easily read in any position.

Disadvantages

Although it can store a lot of characters, far more than most 1D barcodes, Data Matrix codes still have an overall character limit and cannot be read with traditional 1D barcode scanners. It is also not quite as widely support a 2D variant as QR Codes.

Others Random Barcodes

There are a bunch of other visual representations of data that come up in various specialized applications. A few that I've run into are:

• Aztec Codes, which is also ISO/IEC 24778:2008, but which I've only seen used in airlines, as it is standardized by IATA.
• April Tags are a coding system optimized for robotics and visual processing in 3D space. For example, it can quickly determine the orientation of an object based just on the April Tag applied.
• POSTNET. This is the old US Postal Service (USPS) barcode automation. It is a 2D barcode because the length of the bars is used to encode the data.
• Intelligent Mail Barcode is the replacement for POSTNET since 2013. It extends the POSTNET 2-length encoding to a 4-length with a center tracking region.

Third-Party Resources

• Online barcode generator. Generates barcodes in just about every format possible. Can be embedded into a webpage.
• Online label generator. Generates label layouts for things like automotive, nutrition, etc, where a standardized format is available
• GS1 Standards. Global organization that helps coordinate all markings in trade.
• Using Barcodes in Documents - Best Practices. A (somewhat dated) document that goes over a lot of the best practices about printing barcodes, such as minimum size, contrast ratios, etc. especially as it relates to the scanner's resolution.