Henry Petroski. American Scientist. Volume 84, Issue 6. Nov/Dec 1996.
The now-ubiquitous fax machine presents an interesting case study of the importance of context in technological developments. The idea of transmitting written, printed or graphical material over long distances is not new, as demonstrated by such ancient notions as winged messengers and carrier pigeons, and the more modern concept of the postal system. With the discovery of electrical phenomena and the development of such inventions as the telegraph and, later, the telephone, the transmission of words and word pictures by other than an animal or mechanical medium became a reality. Indeed, in the second half of the 19th century telegraphic messages could be carried over transoceanic cables. Telegraph offices served as terminals, which developed into centers to and from which young messengers such as Pittsburgh’s Andrew Carnegie carried the printed documents. A message was encoded, transmitted, decoded and then transcribed to or printed on papers that were delivered to their ultimate destination by other messengers. The advantages of transmitting words in such a way were obvious to everyone, and especially to inventors. It was also obvious that one of the system’s greatest shortcomings was that pictures and graphical materials could not be transmitted directly. Such a limitation of a technology sows the seeds of invention.
In 1843 a British patent for a method to transmit images electrically was granted to a Scots clockmaker named Alexander Bain. Bain was a prolific inventor who perfected the electric clock in 1840 and devised the inked typewriter ribbon in 1841. His inventive interest in both the electric telegraph and the synchronized electric clock prepared him to see how to combine their features into a pair of devices that could send and receive graphical material. As with the early versions of many inventions, Bain’s facsimile machine was crude and clumsy by today’s standards. The graphical material to be transmitted had to be prepared as a raised image on a metal block, much the way the reverse images of letters are raised from typewriter keys. A stylus was passed over the block and rose and fell in conformity with the raised image, thus breaking and making the transmission circuit. A synchronized stylus on the receiving machine was then raised and lowered to reproduce the image in another medium. By incrementally repositioning the transmission block in a second direction, the entire image could in time be reproduced. Bain’s basic idea of scanning an image is preserved in modern fax machines, but the clear disadvantages of his first efforts drove subsequent inventors to seek ways to simplify his large and cumbersome machines and to reduce the effort needed to prepare images for transmission.
Among the first improvements was to replace the bulky, costly and difficult-to-prepare metal plates with tinfoil on which the image to be transmitted was drawn in ink. As the stylus passed systematically over the tinfoil, the current through the circuit was altered, and so the information could be transmitted electronically, to be recreated a line at a time on treated paper by a stylus incorporated into the receiving circuit. The first commercial facsimile system employing such a technique comprised a pair of machines installed in Paris and Lyons in 1865 by Abbe Caselli. No less an inventor than Thomas Edison was interested in improving the new technology, and he attacked the important problem of keeping the sending and receiving styluses in phase and synchronized by employing elaborate gearing and electromagnetic pendulums at each machine, which could be set in synchronized motion at the beginning of each transmission.
One of the limitations of using styluses was that they were principally on-off devices. This was fine for transmitting line drawings and graphs, but it was not effective in communicating the gradations of black and white that are contained in photographs, and so their images could not effectively be sent electronically. Near the end of the l9th century, however, the photoelectric cell was developed, and this enabled photographic images to be scanned and converted into a spectrum of electronic signals, thus transmitting shades of gray. As an added advantage, images could be sent faster, and Arthur Korn, a German inventor, first demonstrated the technique successfully in 1902. The transmission of photographs electronically had an obvious commercial appeal to newspapers, which were thus willing to invest in the new technology, and a wirephoto circuit was established by 1911 to connect offices in Berlin, London and Paris. Networks were established in Europe and the United States after World War I, and transatlantic transmission of photographs was established in the 1920s. In the early 1930s, The New York Times supported the work of the inventor Austin Cooley, who was developing a facsimile machine that was small enough to be carried by an individual to transmit news pictures via ordinary telephone lines.
One of the weak links in the chain of any newspaper-publishing enterprise is the printing and distribution network. The latter is especially capital- and labor-intensive because of the fleet of trucks and army of workers it requires to get the news delivered before it is stale. Newspapers have always been vulnerable to being shut down by strikes of printers and truckers, and a technological means of circumventing that had a clear appeal to publishers. The promise of being able to transmit facsimiles of newspapers around the clock to radio receivers in homes and offices led to the establishment by 1940 of 40 stations sending out experimental newspapers via ultra-highfrequency (UHF) transmission. These experiments were interrupted by World War II, and the emergence of television in the wake of the war led to the abandonment of commercial facsimile broadcasting by newspapers.
Facsimile transmission, albeit relatively crude and slow, was thus well established in specialized uses long before fax machines became the familiar instruments they are today. It is instructive to understand why a century of engineering and technological advances were not alone sufficient to give the system a more widespread presence. After all, typewriters and telephones were also crude by today’s standards, and yet they pervaded places of business of all sizes and in all locations.
Among the reasons fax machines were not more widely incorporated into offices earlier than they were, at least in America, has to do with the communications infrastructure that had developed in the early 20th century. The telephone system consisted of effective monopolies, with the American Telephone and Telegraph Company controlling large networks of phones and phone lines. Customers did not even own their telephones, but rather leased them from AT&T, or what was known as the Bell System, and nothing but the phone company’s instruments could be attached to their lines. In the 1930s AT&T decided not to pursue the development of wirephoto or other facsimile transmission services over its lines, and since AT&T was a monopoly, this natural infrastructure for transmitting data was not readily available to independent inventors or engineers, or to other companies. This is why newspapers used UHF radio waves as a transmitting medium in their earlier experiment.
In the late 1960s, the Federal Communications Commission, which regulates telephone networks in the United States, decided to allow non-Bell System equipment to use the established public switched-telephone network (known in the industry as PSTN). Similar deregulation occurred in the early 1970s in Japan, which meant that facsimile transmission now had an intercontinental communication infrastructure available for its use. Hence, electronics companies and their engineers began to work with renewed interest on the development of improved fax machines that could be linked through the PSTN via acoustic couplers over which the analog data was transmitted much as the voice was.
Competition had its downside, however. Growing numbers of electronics companies, driven by free access to the telephone network, soon introduced a plethora of new equipment that was not readily compatible with the other equipment on the market. This was clearly not a desirable development, for businesses that wished to invest in fax machines wanted to be able to send faxes to every other business with a fax machine; communication between any two independently purchased units was not unlike a return to the earliest fax technology in which machines that were not paired in design did not work.
The dissemination of technology in a free market cannot proceed very effectively if each manufacturer works in total disregard of all the others. Purchasers and users like to be able to mix and match products and components for maximum convenience. When a new technology is being introduced, however, it is not uncommon for different manufacturers to be pursuing similar but different ways of dealing with the technical details. The first light bulbs, for example, did not have standardized bases, but now we expect different brands to be made with the same screw base so that they can be used interchangeably.
Standards for Facsimile Transmission
Fax machines have an obvious need to communicate with one another. A given corporation could specify that all of its offices use the same brand of fax so that communications could flow smoothly among them, but communication across company lines would be impeded if fax machines were not compatible. Such compatibility would come if the different fax manufacturers agreed to common specifications, and the first standards for fax transmission were in place as early as 1968. Since communicating by fax, like telephoning, was an international issue, the United Nations, through its International Telegraph and Telephone Consultative Committee (known by the letters CCITT, because the official name in French is Comite Consultatif International Telegraphique et Telephonique), promulgated the standards, which specified an analog frequencymodulation transmission scheme. According to the 1968 standard, an 81/2 by 11-inch page of information was expected to be transmitted within six minutes with a resolution of 100 dots per inch both horizontally and vertically.
Although technical standards spell out minimum expectations at the time a standard is written, they do not limit the levels of performance that manufacturers can seek. All technological artifacts evolve as new ways are discovered and developed for removing limitations on existing technology. In the case of fax machines, some obvious limitations were the speed and resolution with which faxes could be sent, not to mention the high cost of some of the early machines. As engineers continued to work on such problems to give their company’s fax machines an advantage over the competition, the 1968 standards became too modest. The class of fax machines that met only those standards came to be known as Group 1 machines. In 1976, Group 2 machines were defined as those meeting a more severe standard, one in which analog transmission time at the same 100by-100 resolution was only three minutes.
Social and Cultural Factors
American manufacturers of analog fax machines were not totally pleased with their sales in the late 1970s, and they did not expect that a great deal of capital investment in a new technology would pay off very well. So they chose not to pursue the development of digital fax technology the way the Japanese did. The Japanese were more motivated than Americans to develop digital technology, not just for marketing but for cultural reasons, because the multiplicity of phonetic symbols and ideographic characters in Japanese were not easily coded for transmission via telegraph and related systems.
By the 1970s the world’s largest facsimile research laboratory was well established under the auspices of Japan’s national telephone company. To promote a national fax industry, the Japanese government forced manufacturers to adopt a communications standard that was common for domestic and international telephone operations, funded research and purchased the resulting products. By 1980, the CCITT put forth standards for Group 3 fax machines, in which digital encoding compresses the data to be transmitted, thus lowering the typical transmission time for a page to under a minute. The new standard resolution became 200 dots per inch horizontally and 100 vertically, with a fine-resolution option of 200-by-200 dots per inch. Details regarding session protocol, known as “handshaking,” which establishes how the data are to be trans mitted between two machines, were also included in the new standard, and by the late 1980s convenient digital fax technology was well established and growing at an enormous rate.
Not only were Group 3 fax machines better at universal communication, but they also were easier and more pleasant to use. Digital and computer-chip technology had automated much of the work of establishing the communication protocol between sending and receiving machines, and thermal printing enabled a relatively inexpensive and odor-free paper to replace the expensive and smelly paper that was common in analog machines. These technological developments made fax machines, which frequently had been relegated to mail rooms and operated by designated employees, more acceptable and allowed their use out in the open by everyone in offices everywhere. Indeed, since the newer fax machines, especially the more expensive office models that used plain paper, employed so much technology that was similar to that in a copier machine, in a pinch fax machines could double as convenient copiers.
Among other significant factors that accelerated the acceptance and use of fax machines were some that might be termed extratechnological or social. By the mid-1980s, thanks in part to some very effective television commercials and advertising campaigns, Federal Express had grown into an ubiquitous overnight-delivery service. The term “Fedex,” in fact, came to be used almost generically for all speedy delivery services. In 1984, before every office had its own fax machine, the Federal Express organization invested heavily in a faxing service it called ZapMail, whereby faxed copies of documents could be delivered at unheard-of speeds. Ironically, Federal Express’s promotion of fax technology was so effective in selling the idea of faxing documents that the biggest potential customers bought their own fax machines and made the service unnecessary Federal Express lost over $300 million in the venture, but its enormous success in overnight delivery enabled it to survive even such a large loss.
Another phenomenon of the 1980s that is widely believed to have influenced the rapid adoption of fax technology was the growing dissatisfaction with postal services. People were complaining more and more about how conventional mail (now derisively called “snail mail” by ardent email users) was getting slower and less dependable. Letters were said to be lost and never delivered. At the same time, the presence and use of photocopying machines in offices had made office workers increasingly familiar with the concept of feeding documents into a copier and pressing a few buttons to begin the process. Many newer fax machines were remarkably like office copiers in their looks and operation, so they were less intimidating to use, and there was less resistance to their introduction into the office routine. Faxing a letter was just like long-distance photocopying. By 1987, for the first time, fax machines exported from Japan exceeded that country’s domestic consumption of them. Furthermore, over the course of about a decade, from 1980 to 1992, the cost of digital fax machines dropped by a factor of 30, thus making them affordable by the smallest of businesses and even by individuals.
More than 600 different varieties of fax machines were available to the consumer by the early 1990s. In the United States alone, the market for fax machines grew from half a million in 1985 to 6 million in 1991. By the mid-1990s, personal computers were commonly being equipped with fax boards, which made it unnecessary to print out a computer-generated document before faxing it, or on the receiving end, to have to enter the faxed data into yet another computer. Computer-to-computer faxes thus promised to reduce the amount of paper generated in offices, something neither the personal computer itself nor the fax machine had succeeded in doing. Estimates are that the number of pages transmitted via fax rose from 1.5 billion in 1985 to 17 billion in 1991, an increase of an order of magnitude in only six years. As much as 40 percent of telephone traffic between the United States and Japan was estimated to involve fax transmissions.
Further Developments in Fax Machines Both mechanical and electrical engineers, working individually and in groups on various aspects of the machine, played important roles in the development of the fax from a clumsy and slow analog device that used smelly paper to the compact, fast and user-friendly machines that became as familiar as copier machines and telephones. In the early 1990s, for example, a group of Hitachi researchers and engineers wrote a paper on “compacting technologies for small size personal facsimile,” which appeared in the Transactions on Consumer Electronics of the Institute of Electrical and Electronics Engineers. The six authors, led by a senior investigator whose name, Toyota Honda, suggested more an automotive than a communications engineer, came from five different divisions of the company. The paper reported on two developments that were found necessary to produce a “small size personal machine,” one involving rollers, which is clearly a mechanical engineering problem, and the other involving a computer chip, clearly a electrical engineering problem. Since home faxes are also expected to double as copiers, the quality of the image produced was very important.
Not surprisingly, the research paper of the Hitachi engineers does not reveal many of the details of how they ultimately solved their problems, but this is to be expected in an industry in which success depends so much on having a competitive product advantage. Industrial secrets cannot last long or be effective indefinitely, however. As soon as the new Hitachi compact fax machine was released, competing companies could buy it as easily as consumers could. Engineers learn a lot about different ways of solving problems by taking things apart, an activity that many engineering students engaged in as children and many engineers continue to do throughout their lives. Learning to make a new machine by taking apart something of the competition s is known as reverse engineering, and by such a process it would have been immediately obvious that a big advance in the new Hitachi fax was its reduced number of motors and rollers. Understanding how the quality problem was solved would take a little more work, but in time engineers working for competitors could figure it out or at least understand the problem well enough to come up with an alternative and perhaps better solution, since patented devices, of course, cannot just be copied. It is in such ways that competitive consumer products come to evolve into similar yet different forms. Because manufacturing companies know that their products are never safe from reverse-engineering scrutiny, and that they cannot maintain their competitive advantage for very long, research, development and engineering groups are an essential component of an active industry.
As the size and price of fax machines continued to be reduced, while reliability and quality of image seemed to be constantly improving, the question of speed of transmission remained an issue, at least with some manufacturers and users. Increasingly sophisticated encoding schemes required less data to be transmitted and hence decreased the time it took to scan and send a page. But the limiting factor was the 9,600 bits per second that could be carried over the public telephone networks. In fact, fax technology had become so sophisticated that when a machine detected that it had gotten a noisy telephone connection, successively lower modem rates of 7,200, 4,800 and 2,400 bits per second were automatically employed until one of these fallback speeds proved reliable.
The question of transmission speed led to the introduction of a new standard in 1984, and faxes that fall within it are known as Group 4. In this category, fax machines are designed to communicate over the integrated-services digital network known as ISDN. However, since ISDN itself did not become widely available as soon as originally expected, largely because local telephone networks and switching exchanges were not yet upgraded to ISDN capability, a decade later there were still relatively few Group 4 faxes operating. As with virtually all new technologies or technological advances, the situation attracted the attention of skeptics and nay-sayers and encouraged them to quip that ISDN stood for “It Still Does Nothing” and “I Still Don’t Need.” Whether the letters might ever stand for “I Sure Do Now” depends on what other new technologies might be developed before the ISDN network becomes widespread and before the costs of Group 4 fax machines drop enough to make them fully competitive. By the mid-1990s, there were promising developments that ISDN would indeed become the “Interface Subscribers Definitely Need.”
What new alternative technologies might be developed is not merely a technical question, as the case study of the fax itself makes clear. Even the most technically advanced problems have nontechnical components that affect their solution. Cultural, social, economic and political developments can be limiting factors as much as are the physical laws that govern electronic circuits and mechanical movements.