A grocery question in 1948
The barcode began with a request, not a gadget. According to the commonly told story, the head of a food retail chain approached the Drexel Institute of Technology in Philadelphia and asked whether product information could be captured automatically at the checkout. The dean reportedly turned him down, but a graduate student named Bernard Silver overheard the conversation and mentioned it to his friend Norman Joseph Woodland. The two began working on the problem in 1948.
The details of that first meeting come from later retellings, so treat the exact wording as folklore. What is well documented is that Woodland and Silver filed a patent application in 1949, and that a patent for a classification apparatus and method was granted in 1952.
The bullseye idea
Woodland is often said to have been inspired by Morse code. In the usual account he was sitting on a beach, drew dots and dashes in the sand, and then pulled his fingers down through them so that the marks became thin and thick lines. Whether or not the beach scene happened exactly that way, the idea matches the patent: information stored in the widths of lines.
The 1952 patent, however, showed a circular design, often called the bullseye code. Concentric rings meant a scanner could read the pattern from any direction, which suited the awkward way a package might be pushed across a counter. The system also needed a bright light, a sensor to turn the pattern into an electrical signal, and electronics to interpret it.
That last part was the problem. In the early 1950s the technology was not ready. The light sources were hot and bulky, and the electronics that would decode a pattern quickly and cheaply did not yet exist. The invention was sound on paper but impractical for a store.
Waiting for the technology
For roughly two decades the idea stayed mostly dormant. Meanwhile, several things changed.
Lasers arrived in 1960, and compact gas lasers eventually offered a bright, focused spot that could sweep across a label quickly, which is why the practical effect of laser beams on retail was so large: a scanner no longer had to sit at a precise distance from the label. Integrated circuits made it possible to build decoding logic small and cheap enough for a store. And computers became affordable enough that a retailer could imagine looking up a price in a database each time an item was scanned.
Industrial uses appeared in parallel. In the 1960s, a system of colored stripe labels was tried for identifying railroad cars as they passed a trackside reader, and it was later reported to struggle with dirt and weather. Retail was not the only field that wanted automatic identification.
The grocery industry settles on a standard
By the early 1970s, several equipment makers were working on versions of a retail code. A bullseye-style system was tried in a store trial in the early 1970s, and linear stripe designs were proposed as well. One practical difficulty with the bullseye was printing: if the ink smeared slightly on a printing press, a circular code distorted in a way that made it hard to read, while a set of vertical bars smeared along their length and stayed readable.
Because a code was useful only if every manufacturer and every retailer used the same one, the grocery industry formed a committee to choose a standard. IBM engineer George Laurer is generally credited with developing the linear design that became the Universal Product Code, or UPC, which the committee selected in 1973. Woodland worked at IBM at the time, which is part of why his name is attached to the story, but the final design was a distinct development from his original bullseye.
The UPC’s layout is easy to describe even if the engineering is not. A pattern of bars and spaces encodes a string of digits. The digits identify the manufacturer and the product, and a final check digit helps the scanner catch misreads. The code does not contain the price. The price lives in the store’s computer, which means a retailer can change a price without relabeling every item.
The first supermarket scan
The date most often given for the first retail scan is June 26, 1974, at a supermarket in Troy, Ohio. A pack of chewing gum is the item usually named. The gum has become a museum piece in the telling, and accounts differ slightly on small details, but the broad story is widely accepted: a UPC-marked product was scanned at an ordinary checkout, and the register recognized it.
That first scan did not make barcodes universal overnight. Stores needed scanners and computers, and manufacturers needed to print symbols on their packaging. Early adoption was slow, and some shoppers were wary. A number of consumers worried that prices would no longer be marked on individual items, and debates over that question continued for years. Over time, retailers found benefits in faster checkout, fewer keying errors and better inventory records, and the symbols spread to more types of product.
What the barcode changed
The significance of the barcode lies less in the stripes than in the shared standard. Once every product carried a machine-readable number, stores could track stock, reorder automatically and analyze what sold. The same approach moved beyond groceries to books, parcels, library cards, hospital wristbands and airline boarding passes. Later two-dimensional codes, such as square matrix patterns, store far more data in a small space and can be read by a camera rather than a dedicated laser.
Three points are worth keeping straight when reading about this history:
- The idea (1948 to 1952) and the commercial system (early 1970s) are separate achievements, with different people and organizations involved.
- Credit for “the” barcode is shared among Woodland, Silver, Laurer and the industry committee, and different sources emphasize different people.
- The code and the reading hardware developed on separate tracks. The scanners that read labels at the till have their own history, and their later spread is a story of cheaper optics, as well as the growth of image scanners and sensors more generally.
The simple yet sturdy idea of translating numbers into stripes proved durable. More than fifty years after that first supermarket scan, the descendants of those lines still appear on almost everything sold.