If you were an engineer or scientist preparing technical reports and didn't want a secretary to retype your documents every time you needed a revision or update, IBM's great new word processor, the MTST, was a blessing. Introduced in 1968, the Magnetic Tape Selectric Typewriter, a desk-size word processor, included two cartridges with half-inch magnetic tape, the left one for reading, the right for recording. Each cartridge could store several typical pages, and one could merge information from one tape to the other. Later versions of the MTST used magnetic cards instead of tape reels.

The output "printer" was IBM's revolutionary Selectric typewriter, introduced earlier in the decade, which had a golf-ball-size print head that could be replaced in seconds. A secretary would typically have an array of type balls, which could be switched for different type fonts or special characters for engineering or other disciplines. The typewriter had no moving carriage. So, instead of a heavy carriage moving from right to left as one typed, a light type ball moved from left to right across a page.

Further, the type ball could be raised or lowered half a line space to provide superscripts and subscripts.

The tape cartridges didn't have take-up spools, so each cartridge dumped tape into its own bin. At the end of a session (or at the end of a tape), tape would be rewound onto its own spool.

Some of the circuitry was based on solid logic technology, which IBM had introduced with the 360 computer in 1964. Flip-chip transistors, with three gold-ball terminals, were flow-soldered onto thick-film hybrid circuits on third-inch-square ceramic wafers. But most of the circuitry used discrete components, including electromechanical relays and thyratrons.

The MTST was cumbersome by later standards, but terrific in its day. It was a dramatic improvement over what was available earlier, onerous typing with occasional need for paste-in characters—and retyping for corrections or revisions. It was not designed for general business use.

Along came Evelyn Berezin, with a solution.

In 1951, Berezin left New York University grad school, where she had been working toward a PhD in physics, to join Electronic Computer Corp., which was later purchased by Underwood Corp. to become its Elecom Division (which may have produced the last vacuum-tube computer).

She started as a logic designer and, in time, headed the company's logic design department. When Underwood folded in 1957, possibly because the world was no longer buying vacuum-tube computers, Berezin joined Teleregister, where she designed several online real-time systems, including the world's first online airline-reservation system. That went to United Airlines.

Not quite a dozen years later, in 1968, she left to join Digitronics, a small company founded by a group of senior former-Underwood people. Then she had her brainstorm. Wouldn't it be neat, she felt, to have a word processor that would be useful for everyday documents%#0151;even ordinary letters? Might not lots of secretaries find a word processor even more useful than their everyday typewriter?

An IC-based processor

Just a year later, in 1969, she founded Redactron, the first company devoted exclusively to word processors. Starting from scratch, she was able to use the latest technology. That involved 13 MOS chips, some of them designed by General Instrument, some designed by a young company, Standard Microsystems, and some designed by Redactron itself. One of the chips was actually a very simple microprocessor. The daddy of commercial microprocessors, Intel's 4004, was introduced in 1971, the year Redactron's word processor started shipping.

In an early design, the machine used small reels of magnetic tape to store files, as well as magnetic cards, which could store about a page per card. It was not possible to edit from page to page, a limitation that could have been eliminated if the machine had access to central memory in a large computer. Later on, when floppy disks hit the market, Redactron used them for file storage. A ROM chip stored the program.

The system occupied about 1 foot X 1 foot X 2 feet, most of that space devoted to the power supply. It had a separate keyboard and printer. One board, about 12 inches square, included all the circuitry. Later machines, which included monitors, could do most of the important things—like arranging columns and margins and moving copy around—that were routine in later-day word processors.

The selling price of the first system was about $8,000, and some machines were sold to OEMs like Sperry Rand. But the industry at the time was accustomed to renting and leasing, so Redactron had to develop a rental program to finance customer loans. That called for a strong bank account, which was not one of the company's strengths. It did have a good relationship with the Bank of Boston, which helped, but there was a great deal of financial pressure on the young company.

Though the company was very small, it was the second largest word-processing manufacturer in the world, after IBM. It set up an international division and shipped quite a few machines abroad—to Germany, England, South Africa and Hong Kong.

The impact of money

Then came the crunch. In 1974, frightened by high inflation and the specter of inflation, the government instituted wage and price controls and cut back on the amount of money that bankers could lend. Redactron's ability to borrow money was severely restricted. It already had outstanding loans because of its rental business.

The financial pressure was too much. In 1976, with revenue of about $20 million, most of it rental income, Redactron had no choice but to sell to a larger company with stronger financial resources.

Berezin envisioned the day when general-purpose computers would become word processors and she wanted to tie the word-processing business, which she regarded as essentially a secretarial business, into the computer industry.

Berezin sold the company to Burroughs. She knew lots of people there, and the CEO, Ray MacDonald, a very astute businessman and marketing man, saw the potential of Redactron's word processor.

The next battle

Unfortunately, the people who worked there didn't. Berezin thought that her job was to integrate her machine and her concept into computers. She prepared a plan for converting her word processor into a terminal for Burroughs computer systems. If her machine could access the memory in a large central computer, which could easily accommodate it, the word processor's capabilities could have been expanded greatly. Berezin expected that every secretary could have a terminal.

But Berezin suffered from a major problem, a not uncommon defect. She was an outsider. The inside people were unconvinced. They felt they could write a small word-processing program that would convert their standard data-processing terminals into word processors. They said they had one standard terminal and they wanted to stick with it. But the standard terminals of the time were quite unsuited to typing applications.

They bought a rudimentary word-processing program that one of their customers had written and installed it. But the terminals were much too slow. You could indeed type on the keyboard, and what you typed would indeed appear on the monitor screen—after a while. There could be a delay of seconds from the time you struck a key till the character appeared on the screen. One wag suggested that you could type a letter, go out for coffee, then return to find your letter on the screen.

Berezin fought for her concept but, she says, there was no way she could overcome the powerfully entrenched Not Invented Here syndrome. She left Burroughs around the end of the decade. She started and ran a venture-capital organization, which she sold after a few years. She now serves on the boards of directors of several companies.

The people inside Burroughs killed her word processor. But it probably didn't matter. Personal computers and a wide variety of inexpensive word-processor software packages arrived in the early 1980s. They put word processors in the hands of ordinary consumers, not just secretaries. Those word-processor packages would have killed Berezin's word processor anyway.

But Burroughs might have been there first.

The Century of the Engineer: Misunderstood Milestones