Telecommunication in the News Industry: The Newsroom Before and After Computers
WILSON R. LOCKE
The newsroom at the Los Angeles Times is supported by one of the largest news editing systems in the world. The computer system is unique in that writers, editors, and others with a news background designed it and are now responsible for its management and development. The successful integration of electronic technology into the manual world of the newsroom, illustrated in this case study, is attributed to extensive staff involvement, cooperation across departments, simultaneous development of the appropriate technology, and management's commitment of time and resources.
In the days before computers, video display terminals, and downloads, there was a primitive instrument in the city rooms of America known as the typewriter: slow, noisy, and unreliable. Often it ran out of a thing called a ribbon, jammed, and always required excessive manual effort. The typewriter was a primitive tool, often battered beyond recognition by reporters and editors venting their hostilities by using excessive force on the keyboard. It was appropriate for a place where reporters smoked, photographers kept bottles of alcoholic beverages in their darkrooms, and editors were forever spitting in the wastebaskets—a practice not only condoned but encouraged, since it kept them from spitting on the reporters.
This was a terrible time in the history of American journalism,
and I am pleased to report that times have changed. Today we are blessed with computers that are both silent and efficient. They have helped change the newsroom from a noisy, smoked-filled hovel into a quiet, air-conditioned room with indirect lighting. The new systems are not without problems. The technology is accused of causing blindness, sterility, and carpal tunnel syndrome. People still batter keyboards, are not above spitting in wastebaskets, and vent their hostilities by sending electronic messages to each other. Yet the physical changes are suggestive of more fundamental changes in decreased costs, improved communications, and greater flexibility. These three elements have been critical to the newspaper industry's ability to benefit from entirely new methods of information gathering and dissemination brought about by technology.
At the Los Angeles Times, the Editorial Computer System, shown in Figure 1, is the core of a network of computers that supports more than 1,200 terminals and more than 1,500 people. It is used to write, edit, and publish news stories and to communicate with a staff around the world. Each day, as many as 25,000 messages move through the network and hundreds of stories are written, sent, and received. More than 100,000 stories are stored on the system.
The network is composed of 24 Tandem computers in four systems. The main system and a small portable one—called the road show—are located in Los Angeles. The others are located in a plant in Orange County (40 miles south of Los Angeles), and in the Times' bureau in Washington, D.C.
The main system supports many local news bureaus, receives incoming stories from news agencies, backs up the Orange County system data base, and is used to produce copy for daily publication. The road-show system is used to cover news events outside of Los Angeles, such as political conventions and the Olympic games, and to test and develop new programs.
The system in Orange County is a smaller version of the one in Los Angeles and, like the L.A. system, supports a host of local news bureaus and is used to produce type for publication in the newspaper. This system also serves as a backup for the L.A. system and can operate independently, duplicating incoming wire service stories, communicating with the ''outside world,'' and composing stories for publication.
The Washington, D.C., system supports the news bureau in that city and acts as the communications center for all Times bureaus outside of California, including those in other countries.
Although the D.C. system is capable of setting type, it is not used for that purpose because there are no production facilities or typesetters at the bureau.
The technological change was not achieved easily. The motivation and the process required for this successful project demonstrate the importance of planning, staff involvement, patience, and resources.
In 1974, managers in the Times' Editorial Department identified three reasons for considering computers in the newsroom: economics, communication, and flexibility. Computers offered the potential of a more efficient way to produce the newspaper by reducing costs and speeding up production at deadline, thus maintaining a competitive edge while also increasing editorial control over the production process. Computers could facilitate rapid communications among writers and editors as well as intelligent, high-speed connections to the larger world of developing information technology. Finally, computer technology offered alternatives in case of emergencies such as power failures in a geographically diverse organization.
Historically, the number of dollars given over to the correction of type is almost beyond comprehension. But more meaningful to the news department have been the hours upon endless hours spent waiting for type to be reset, and the drudgery of overseeing it replaced in the paper. This chronic necessity for correction, it should be said, did not stem from lack of diligence or competence in one department or another. It is traceable to a shared accountability between two departments (Editorial and Production) and a score of intervening nuances, right down to the age of brass mats in Linotype machines.
One potential saving of a computer system was to avoid retyping stories. The buzzwords were "save the original key stroke." Rekeyboarding stories introduced errors that cost thousands of dollars in proofreading and corrections. If news type was more accurate it would lead to smoother production and a decline in the number, and thus the cost, of corrections. Also, the increasingly insistent production demands had to be more readily accommodated.
As part of reducing costs and improving efficiency, an editorial system would also lessen the problem of shared accountability by making those who write and edit words also set them into type. The day the first scribe laid down his quill pen in behalf of flexible type, he unknowingly put into the hands of others the responsibility for validating what he had written. The by-product of this transition was not apparent in the measured pace of early typesetting. However, as the tempo of newspaper printing increased, there arose among writers and editors a sense of frustration in converting their words faithfully into type. The computer would return the English language to the control of the Editorial Department and, at the same stroke, place in the department's hands a tool to vouch for the integrity of type, a prospect that gave its users new inspiration.
Just as the integrity of a reporter's thought and research should be preserved from its original keystroke, stories from reporters based outside Los Angeles must move to the central office with speed and reliability. Experiments with portable terminals already demonstrated the benefits to be accrued from technology. Stories were transmitted quickly from any telephone anywhere in the world.
Terminals in all local and national bureaus, as well as portable terminals for those traveling on assignment, brought dramatic changes in news gathering for the Times. Before these machines were introduced, news copy was delivered in a variety of ways: hand carried by messengers, by dispatch truck from suburban offices, by facsimile machines, by telephone dictation, by Telex, by mail, and in one instance by carrier pigeon. In each case the time between completion of a story by a writer and delivery to an editor was measured in hours and sometimes days. This had the sinister effect of sacrificing the decision-making process to the demands of setting type.
Electronic technology, as it was practiced in the editorial bureaus, reduced this writer-to-editor segment to minutes. When necessary, a writer could move a story to his editor in installments, transmitting his lead and succeeding parts to Los Angeles, while simultaneously writing the remainder of the story. A news editing system could refine this capability, increase the speed of transmission (making two-way communications more feasible), and provide more time for thoughtful writing and editing.
The facility in Times Mirror Square is a complex structure made efficient by centralization of elements required to produce the daily newspaper. In recent years satellite plants and bureaus outside downtown Los Angeles expanded the newspaper's production and news-gathering capabilities. At the same time, the communications required for this diversification created a vulnerability to the need for continuous electrical power and telephone service. In case of a disaster, publication would be extremely difficult because of space limitation in any one Times plant, the problems of moving people and equipment over difficult terrain, and the loss of essential information. In the instance of failure in one location, a more extensive and integrated computer system would provide backup capabilities in another location.
The creation of the Times' news editing system started in 1974 when the decision was made to explore electronic technology. Because the Data Processing staff did not seem to understand the needs of the Editorial Department, management decided to use editorial staff to develop the system. A research team was formed. It consisted of two news editors and a composing room foreman. A "fourth member" of the team was one of seven writers, editors, and clerks who worked with them on a two-week rotation. This group made all recommendations and basic system decisions. A larger advisory group of 23 staffers also was formed. This group met once or twice a month with the research team, to apprise each other of progress on the project and changes in the news room. Their efforts were coordinated by an assistant managing editor who kept senior management informed, approved expenses, arbitrated disputes, and kept the project on course.
In the seven-year odyssey that followed, the staff learned a new language of bits, bytes, baud, and buffers; they learned the difference between needs and wants; found out how much they did not know about their own operations; and established department objectives. The process was divided into three distinct parts: planning, development, and implementation.
The first year was spent orienting the editorial staff and defining needs. A small editing system consisting of a computer and
two terminals was leased from the Hendrix Company. In a period of six weeks, more than 400 members of the Editorial Department—in groups of fours and fives—were shown how an electronic system operates. After the demonstration, each participant prepared a detailed job description and discussed with the team the problems faced every day.
Specifically, each person wanted more control over the words written and the words edited. In the wider view, the staff asked for improvements in the amount of time it takes to get a news story into the office and set into type. Uniformly, all said they wanted new tools to be fast, accurate, and dependable. Based on these discussions, the following six major objectives were identified:
Maintain editorial integrity over the written word and its translation into type.
Produce a near-perfect newspaper at an hour more acceptable to the Editorial Department, but simultaneously more acceptable to the Production and Circulation departments.
Receive, route, write, and edit news stories with flexibility and at a speed compatible with the anticipated changes in typesetting equipment.
Create an indefinite memory of news and an indexing capability responsive to any reporter or editor's command.
Establish standing formats for tabular copy, such as earnings reports and box scores, which could be called up on a terminal, filled in, and released to type upon receipt of the data in the Financial or Sports departments.
Dramatically speed up news communications, whether from the Civic Center office across the street or the Times bureau in Moscow.
The second year was devoted to studying the existing production methods, preparing specifications for an editing system, and surveying how other newspapers were using electronic technology. Copy flow charts were prepared for every department, and conferences were held with the Production, Information Systems, and Telecommunications departments.
Particularly valuable were interviews with editors, writers, and production people at more than 15 newspapers around the country that already had electronic newsrooms, for example, the Louisville Courier-Journal , the Tucson Citizen, the Houston Post, the Vancouver Province and Sun, the New York Times, and the Long Beach Press Telegram. Criteria for these interviews included the users' experience and evaluation of their system; their evaluation
of the vendor's competence, size, and stability; how their system software worked; and if the vendor seemed to understand their system objectives and requirements.
Universally, the data processing people and technicians were happy with their system; editors and writers were either very proud of it or dissatisfied because of the amount of computer coding required to do their jobs. Valuable insight was gained into the relationship between buyers and sellers (ranging from good to poor, but improving) and what could be expected after contracts were signed. More important, it appeared that no system on the market would come close to accomplishing what was needed at the Times. A new system would have to be developed.
Study of the Times' production methods continued in the third year, along with rewriting the specifications, conferring with vendors, and preparing a return-on-investment statement that justified the cost of an electronic system. A major decision point that emerged was whether to do the project in-house on an existing, centralized mainframe computer or to hire an outside vendor and use minicomputers in a distributed system.
A software house was commissioned to study and report on the options. They sent a team to Los Angeles to spend a month interviewing editorial, data processing, and production people, reading the specifications, and probing the operations. The report said that neither a centralized nor a distributed approach was inherently superior. Different aspects of each system seemed best suited to specific requirements.
The main item of controversy was whether the mainframe computer could handle the increased work load and at what cost. The Data Processing Department wanted to use the existing computer to implement the editorial system and argued that the existing computer was underused. In addition, the Data Processing Department planned to install a second computer, primarily for another department, which they said could also handle the sudden impact of the greatly expanded computing requirements. In the Data Processing Department's view, the mainframe would easily accommodate the local terminals, as well as those in Orange County and the bureaus. They argued that a centralized system was also feasible since the news editing system requirements were more straightforward than those encountered in most applications because the editorial sub-departments operated on similar information. The Data Processing Department approach would create a separate file management system for Orange County on the computers in Los Angeles.
Editorial argued there would be no local processing capability in Orange County, and that the plant would be totally dependent on Los Angeles for its operation. Editorial wanted a system in the Orange County plant that would be configured with a smaller subset of the Los Angeles system, use identical software, and be able to operate independently.
The report by the software house went on to address the Editorial Department's desire to program, maintain, and control the news editing system. It said that implementation of either system was likely to result in split leadership in the area of data processing. It warned that "If, indeed, a situation involving split leadership should remain after system completion, it would take an integrated system impossible, seriously impacting such long-range company goals as pagination."
The evolution of information systems also provided some insight into the advantages of functionally specialized distributed systems. After several years of frustration with attempts at centralized systems, the trend changed, and the predominant model became a set of information systems modules, each designed to accomplish a relatively small task. Thus, the systems are designed proceeding from the small to the large, not conversely.
The success of the modular systems elsewhere has been attributed to several factors. First, the information needs of a single department are better defined. Second, when the system is developed in the department where it is to be used, it is natural to involve the ultimate users in the design decisions. Thus, cooperation is fostered from the beginning, and the system is tailored to its environment. Third, since the job is more circumscribed, the system can be brought into use and tested relatively early, while there is still enthusiasm. Thus, the system can evolve in use. Fourth, it is more practical to arrange for backup of a distributed system. Finally, new systems can be developed without affecting those in regular operation—which is simply not true of centralized systems.
The report concluded that the following substantial advantages were to be gained from a distributed design:
Inexpensive and effective backup could be easily provided.
Pieces of the system could be easily and safely modified, developed and tested in virtual isolation from the production system.
Functional specialization could easily be achieved, for example, special facilities provided for the sports desk or the editorial page.
Users' involvement in the ongoing implementation could be encouraged and simplified.
Managers in the Editorial Department decided to develop a distributed system along these lines recommended in the report:
Prepare a management plan for a news system and provide a clear definition of system goals. Develop milestones, establish areas of responsibility and concepts of operation, and set future goals for pagination and backup.
Revise the specification to address a Phase I news editing system, including specific design criteria and building blocks for the full system, particularly full editing capabilities, communications input, printer subsystem, and basic story file maintenance.
Initiate a study of story file and directory-handling requirements to enable data base management functions and performance criteria to be specified for Phase II.
Issue a specification for Phase II development after a high degree of confidence is attained in the performance of the Phase I building blocks, so that full system functions can be specified.
The report laid a solid foundation and brought the project into focus. The third year ended as it had started, but with renewed staff vigor and a new perspective.
In the fourth year a return-on-investment (ROI) statement was completed and a contract signed to build a news editing system. The ROI statement listed all savings, from paper clips to long distance telephone toll charges, and all costs, from programmers to taxes. It showed a 12.5 percent return and a payback of approximately six years and four months on capital expenditure. The need to move forward was reinforced by the Times' growth, a vastly expanded program of news coverage, increasingly stringent deadlines, and growing press runs. The conversion in 1974 to photocomposition (abandoning hot metal in favor of film), required the Editorial Department to execute an increasing number of intricate procedures in a narrower time span. Photocomposition in particular, and the knowledge gained by talking to other newspapers, made it clear that the Editorial Department could and would begin to set its own type.
A standing joke in the newsroom was to ask people working on the project when they would have a system, and the stock reply was "next year." This did not change in the next few years.
Years five and six were spent building the system and refining requirements (finding out what was possible and what was not). Several major events occurred after the contract to build the system was signed.
The first event was that the Data Processing, Editorial, Production, and Telecommunications departments, following months of vigorous debate, united with a common goal. Task forces were formed to attack the problems, and weekly conferences kept each team abreast of what was happening with the others. One group worked on the terminal, another on printers and communications, another on the data base and user applications, and another on typesetting. Each group brought to the effort an impressive expertise.
An implementation schedule was drafted, discussed, and adopted. An acceptance test plan was devised. Acceptance tests were to be performed at established milestones in each stage of development, with a final performance test when the system was completed.
The next event was an alliance of "systems" people at the Times and the other newspapers we had visited. When applications questions arose, we started contacting the experts at those newspapers. For example, to resolve typesetting questions before a test system was in place, electronic copy was sent to staff at the Vancouver Sun or the San Francisco Chronicle to be set into type on their systems and returned by mail. Soon they were calling back to ask how we were handling a problem they faced. This group shared insights, exchanged information, and developed new concepts. This alliance has continued through the years and become a valuable resource in the daily battle to make things work.
The third event that had an impact on the project was the confederacy formed between vendors and the alliance of systems people. The annual American Newspaper Publishers Association Technical (ANPA/Tec) conference brought together newspaper people and most of the major vendors of newspaper equipment.
ANPA committee meetings, such as the one on wire service communications, were attended by vendors and systems people. In 1977 the wire service committee was struggling with the transition from teletypesetting (TTS paper tape) to electronic transmission of copy and the coding to allow computers to recognize and interpret it. It was a complex problem because one bit of information out of alignment would turn a fine news story into garbled trash. The main objective of the wire committee was to establish communications standards.
Systems users and system vendors talked to each other about
standards and solutions to mutual problems. This dialogue resulted in systems, using different computers, being programmed to communicate with each other or at least being able to transmit to each other. It also opened the door for vendors to discuss and resolve problems concerning word hyphenation and to develop mutually satisfactory applications.
It was not uncommon for the people who were building the Times' news editing system to discuss a problem with established, competing vendors. A difficulty with this candor and openness, however, was that innovations would lead to renegotiating the system being built. This and the complexity of the job led to delays, missed deadlines, and cost overruns in the project.
Building the system was slow and complicated. As programs were written, tested, and revised, the blocks fell into place. The terminal functions for typing, scrolling, defining and moving copy, inserting, and editing were progressing; the printing devices were being driven by the main computer; the typesetting programs were nearing completion—kerning tables, hyphenation algorithms and type-font tables were being refined; the communications programs were receiving copy from Associated Press and United Press International; and the data base programs were beginning to come together. As each segment was completed it was integrated into the main system.
Then we were plagued with a rash of delays. The typesetting program was not working, and an expert was hired to resolve the problems. Next a glitch developed in interfacing the terminals. Every time one fire was put out, another would start. In the latter part of the sixth year, amid feelings of despair and desperation, it was no longer funny when someone asked when they would have a system. The publisher, managers, and editorial staff were losing confidence and interest in the project. Then, by mutual agreement, the project was abandoned because of the delays and mounting costs.
At the same time, however, a second generation of systems had arrived. The search was resumed. The Washington Post had embarked on a similar project and that system was studied. They had a beautiful terminal with a very readable screen display. The computer met the Times' criteria for a redundant system (in which every element has a backup), and the applications software was first rate.
Although the Post's system was still under development, it was the first system that addressed the needs of a large newspaper (400 plus terminals) without compromising performance and func-
tionality. Concern about the time the system took to process requests was resolved. Changes made to the communication software, however, forced the Times' team, reluctantly, to reject the system because the change would have made the required communications links to Orange County unworkable.
During this period information was gathered on ergonomic issues such as lighting, radiation, seating, keyboards, and workstations. The newsroom was rebuilt to accommodate the new technology, anticipating the future.
While the staff had been working on a unique system for the Times, an off-the-shelf product was developed that met 99 percent of the requirements. It was real and it worked. In 1981 the Editorial Department purchased and installed a small development system that was to grow into the current sophisticated news editing system.
The Editorial Department would have been naive to undertake these new responsibilities with anything less than great care. Nothing was more sobering than the contemplation that we were to have a key role in the very complicated process of setting type. The editorial staff sought and received counsel and help from the Production, Information Systems, and Corporate Management Services departments. A phased approach was used, starting small and building block by block.
After the news editing system was installed, there were clear signs of success. The network united writers and editors around the globe, and there were immediate savings in production. Failures were few and, for the most part, insignificant. Goals were achieved in terms of economies, communication, and flexibility.
Typesetting savings included better ability to cope with the late changes in stories and late advertising changes that force revisions in page layout. Lost or misplaced type and the need for setting and storing large amounts of type before pages are ready for makeup were reduced. In addition, there was a means to check the proper placement of the elements of composition such as bylines, subheads, windows, box inserts, and jump lines.
Writers and editors gained almost instant communications with one another, whether they were in the same building or separated by a continent. Wire service stories, as well as stories and messages from staff writers and editors in remote locations, were automatically sorted as they arrived and were delivered to the ap-
propriate department, desk, or person. An unlimited number of editors and writers had simultaneous access to stories coming into the system, thereby eliminating a sometimes long waiting period for delivery of hard copy from the wire room. Appropriate editors were notified immediately of wire service bulletins or urgent copy through special ''alert'' software provided with the system.
The system constantly monitors communications lines in the Times' network. Under normal conditions the system sends stories and messages to the various Times' offices over preselected routes. However, if a line is disabled, the system will automatically reroute the transmission over the next best path, even if that includes sending it through another sex of computers in Orange County or Washington. Through a process of software searching, the system also scans incoming stories for specific words and from this makes topical compilations of stories into files such as Economics, Oil, Space.
For the Editorial Department, the sophisticated communications abilities of the system make feasible such flexible alternatives as operating from the Orange County plant in a backup mode. The computers' communications program and the Times' fiberoptic link with Orange County unite two elements of the news editing system. Coupled with satellite dishes in Orange County for wire service reception, writers and editors in either Los Angeles or Orange County can produce all or part of the editorial portion of the daily newspaper. If, for example, the Los Angeles production facilities were inoperative but the news editing system was intact, this software would allow the Editorial Department to write and edit in Los Angeles while setting type in Orange County.
The news editing system also extends the reach of either operating plant to encompass other facilities, bureau offices, temporary offices, or mobile stations—any location where at least one terminal and a telephone can be set up. In effect, one telephone call turns Ventura or Riverside or San Diego into one corner of the newsroom. As an extreme illustration, in an emergency editors, copy desks, and production facilities could be set up in the Orange County plant while most of the reporting staff operate from suburban bureau offices closest to their homes or in their homes using personal computers.
Finally, the news editing system becomes its own savior, helping to maintain its own integrity as well as that of the daily and standing news file in Los Angeles. As stories reach critical bas-
kets in their evolution, the system automatically, or at a staffer's direction, sends a copy to another system for safekeeping. Magnetic tape is also used for storing backup copies of the text and system files.
In either case, essential information and the foundation of a news file are available in duplicate and in widely separated locations. With its speed and communications intelligence, the news editing system is a positive response to single-point failure and, in spreading the Times' resources, eases the threat of vulnerability.
Staff acceptance of the news editing system was almost total with very little resistance, although there are still several people who would rather use pencil and paper. Comments from writers and editors illustrate the positive impact of the system. For writer Martin Bernheimer there were professional benefits once he overcame his resistance:
I was the last convert. I didn't want to join the 20th century. I liked torturing my prose to death on paper. I enjoyed rewriting each sentence a dozen times and driving typesetters crazy with penciled overscores, with special instructions scribbled in margins and with arrows connecting sentences a mile apart. I gave up my battered friend only with a fierce struggle, and some gentle, patient coaxing. Who needed these fancy new-fangled contraptions anyway?
I did. It took me about three hours to discover the joys of clarity and flexibility as the contraption actually made me work better, not to mention faster. Contrary to expectation, the computer did not make me a slave to technology. It may not have made me free, but it certainly did make me a more efficient, more professional, more accessible writer. My trusty new Coyote (the name given to the terminal) has changed my professional life—and changed it for the better.
Editor David Blume wrote about professional and personal benefits beyond speed and efficiency:
In terms of general use, of course the Coyote when it arrived provided speed and efficiency. I am much swifter now that I ever was at a typewriter, even though I am not a formal touch typist. But in my case, to use it in an ordinary way was insufficient and noncreative. At first, the method generated more interest than substance. The ability to create shortcuts became an end in itself, and I spent every idle moment creating new user keys. About a third of them turned out to be extremely useful. The rest were ideas that weren't so good in execution as
they were on the drawing board. Usually, finding out where I went wrong was a humbling experience. The mistakes were almost universally stupid ones.
Another side effect was the interest working on the Coyote generated in electronic music devices. Since I started using the word processor, I have bought six electronic keyboards, including the latest, a very intricate sampling machine. All of them have required a great deal of homework and practice, but the rewards are substantial in terms of creative results at increased speed.
There were five key elements in the successful implementation of the system:
Constant interaction with the staff, from clerks to editors
Resisting the temptation (and at times it was tempting) to select a system that did not meet our needs
Development of appropriate technology
An enlightened management that allowed time and funds to achieve the objective, with some mistakes along the way
The news editing system also introduced a language expressed in digital form. This digital transformation makes available an entirely new field of information gathering and dissemination. Foremost among the prospects are a computer-based library system, pagination, and exchange of stories and information with such organizations as wire and weather services, government agencies, and cable television.
In the short term, the typeset version of a news story is also the preliminary library version. The Editorial and Information Systems departments are working on a program of full-text retrieval. This program allows librarians to call up stories through the news editing system for indexing and enhancement in order to make these stories available to editors and writers through a central data base. It also provides an information bank of Times stories that are a marketable commodity. Repackaging of news for distribution by such agencies as Times Mirror Cable is also an obvious consideration.
The next large step in news technology is news pagination, which involves the electronic composition of news pages. This includes text, ads, graphics, and pictures as well as integration of all production aspects from editing to distribution and billing. It
is practiced today by several newspapers, and while the Times has pioneered in display ad layout, full-page classified ad and stock page makeup, this competence is only a first step.
The next generation, or fourth wave, holds promise of even more sophisticated equipment, such as voice recognition computers, and facsimile devices that convert documents into text files. This equipment will allow us to communicate and produce our newspaper with even greater speed and accuracy.
Many pieces of the puzzle are in place and working. The task is integrating all the pieces into one homogeneous unit that satisfies the requirements of the Times and, especially, the people who produce the newspaper.