This is a modest (one might say half-hearted) attempt to give some context for where the Wang 1200 stood. The emphasis is strictly on Wang, with a brief mention only of IBM's offering. Although there were certainly other products in the market, Wang saw themselves competing primarily with IBM.
Wang Laboratories was founded in 1951 by Dr. An Wang, an immigrant from China. Wang had gotten Masters and Ph.D. degrees at Harvard, where he also made some key inventions in the development of core memory technology. Some of Wang's products were brought about by his own determination to create it; others were done under contract. Wang grew slowly at first, and could support only a small staff. That changed with his first big contract.
Compugraphic (link)
In 1962,[1] Wang was approached by Compugraphic, another small New England company, one that specialized in phototypesetting equipment. They had an idea for a relatively inexpensive ($30K vs $1M) computerized phototypesetter.
Historically, typesetting had been done via metal type face that was assembled manually at the hands of an experienced typesetter, one that could anticipate how much leading to place between characters and words to produce nicely justified text for books and newspapers.
Phototypesetting equipment used optical means, not lead type, to produce the image of each page to be printed. Phototypesetting had the same difficulties in justifying the text placement, but it was replacing the hot lead method, as the results were of superior quality and could also be cheaper.
As the price of electronic computers came down, the algorithms to space text on a line and break at word boundaries became in reach. The last stumbling block, though, was hyphenation. Hyphenation is required to get acceptable justification, which meant either a very large dictionary of words with hyphenation, or a very complicated rule set combined with a smaller dictionary of exceptions. Although some very expensive ($1M) computer-drive phototypesetting systems were available that could do this, the price tag prevented all but the largest newspapers from using such a system.
Compugraphic's idea was to strike a balance. A simple computer would be used to automate the spacing of letters and words, but any time it needed to hyphenate a word, the machine would stop and a red lamp would signal to an operator that help was needed. One operator could keep five or so Compugraphic machines running at near capacity, at a price lower than one fully automated phototypesetter.
Compugraphic didn't have the ability to design such a machine, so they gave the contract to Wang Labs to design and build one. In under a year, with a total investment of about three man-years of effort,[2] Wang produced the Linasec semi-automatic phototypesetter. Each machine sold for about $30,000. In 1963 Wang sold about $300,000 worth; $470,000 in 1964, and $640,000 in 1965. It was the dominant source of revenue of Wang Labs in those years.[3]
This profitable arrangement came to an end around 1966. The way Dr. Wang tells the story in his autobiography, 'Lessons,' Wang took the financial risk of developing the machine and owned the patent on it, and Compugraphic gave verbal indications it wouldn't compete against Wang, although they retained the right of manufacture.[4]
Compugraphic executives tell a different story. They specified the machine and Wang did an admirable job producing the machine. At first both enjoyed the profits of the growing business, but when it plateaued, Compugraphic argued that Wang had to settle for a smaller profit on each machine to grow the market, but Wang wouldn't hear of it. Compugraphic started making the machine themselves despite Wang threatening to sue them on the basis of his patent on the machine, a patent which was a surprise to Compugraphic.[5]
Compugraphic later went on to design other, more capable machines, and was prominent in the phototypesetting industry even twenty years later, in 1989, when they merged and became part of Agfa.
Wang 1220 (link)
Following the collapse of the revenue stream from the Linasec machines, Wang successfully switched over to selling a series of innovative and increasingly capable electronic calculators. The pinnacle came with the Wang 700 family and its derivatives, such as the cost reduced Wang 500, but Dr. Wang saw the writing on the wall, and knew he had to find a way out of the calculator business before it became a commodity.
Wang decided to tackle two different markets. The first was to get into the computer business, but that lineage (Wang 4000, 3300, 2200, VS, PC) won't be pursued here. The other was to get into office automation.
IBM started selling their Selectric typewriter in 1961. Only three years later, in 1964, they started selling the "MT/ST" (Magnetic Tape Selectric Typewriter). The Selectric was augmented with solenoids for activating the printing mechanism under electrical control. A very small computer used two spools of ½" wide tape, each tape storing about 25,000 characters, to do simple word processing tasks. Later, in 1967, IBM changed to using magnetic cards instead of tapes.
Wang rented an IBM MT/ST, analyzed it, and decided they could do better. The Wang 500 calculator already had the ability to use an OEM'd Selectric as the input and printing device, and it also already used data cassette drives of Wang's own design to store programs and data. The only thing missing was the microcode to make it work as a word processor instead of a calculator.[6]
Harold Koplow, who had won an internal competition and earned the job of writing the Wang 700 microcode, was given the task of rewriting it for the Wang 1200.[7] The product decision was made in 1970, and by November 1971, Wang announced the Wang 1200.[8]
Wang started shipping the first model, called the 1220, in May, 1972.[9] Although introduced with high expectations from Wang, early units of the 1220 had difficulties in getting consistent results, as the carriage didn't keep an even path across the page. Up to 80% of the contracts for 1220s were canceled.[10] After months of investigation, it was discovered that IBM's OEM'd Selectrics were missing a certain spring which stabilized the carriage, while all the Selectrics sold by IBM had this spring. Despite IBM's denials that it was required, IBM came out and retrofitted all the OEM'd Selectrics.[11]
Harking back to Wang's early experience building the Compugraphic/Linasec system, the 1220 was also offered with a photocomposition option,[12] namely some extra commands and a link to a paper tape punch. The Wang 1200 operator could edit and revise copy on the Selectric, using the Selectric printer for proofing, then when the copy was complete, hit a key sequence to produce a paper tape suitable to drive a phototypesetting machine. The text stream could embed control codes to allow specifying changing of type face and font as well.
The 1220 could also take a telecommunication option, essentially a serial link to an external computer.
Apparently, the early 1220 units used ferrite transformer ROM technology to hold the microcode. At 2K words of 43 bits, it was fairly expensive. This was later replaced with mask programmed MOS ROMs. Although the backplane was unchanged, a few cards were revised (6170,6171,6172,6173 → 6370,6371,6372,6373).
Wang 1222 (link)
In January 1975 Wang added the 1222 to the product line.[13] The 1222 added a one-line CRT so the line being edited could be viewed on-screen before being committed to paper.
The hardware was a mix of old and new. The 6370, 6371, 6372, 6373 boards were replaced by 6716, 6717, 6718, 6719, 6720. These are described as the Typewriter Input, Teletypewriter output Driver, Teletypewriter Output control, and KA KB Code Converter. The 6718 contained an Intel 4004 four-bit microprocessor, along with 1 KB of EPROM code. The MOS ROM card, which held the microcode for the bit serial CPU, was still 2Kx44, but there was now 12 words of discrete diode ROM override to patch the mask ROMs.
As an example of how relatively sophisticated these systems could be, an issue of a Wang newsletter had a small profile on Argus Research. Argus had five 1220s in operation, with the photocomposition option, a Compugraphic Computape II phototypesetter, and a Compugraphic Uniterm CRT paper tape editor.[14]
Ultimately, sales of the 1200 product were not what Dr. Wang had been hoping for, although it wasn't abysmal. Rental and sales of the word processing equipment accounted for 10% of the annual revenue for the year ended June 30, 1976.[15] This end of year report coincided with what was one of Wang's greatest product introductions.
One book pegged the total number of Wang 1200 family machines sold at 3000.[16]
The Birth of WPS (link)
In an oft-told story, Harold Koplow and Dave Moros, very important engineers in the late 60's and early 70s, were on the outs with the now much larger 1975 Wang. They had been relegated to long-range planning, a presumed easy transition out of the company.
Instead, they did something that saved the company. Koplow had been the one who wrote the microcode for the Wang 1200 based on the modified Wang 500 calculator hardware. Guided by Ed Lesnick, Koplow and Moros took some trips to various business who were then creating documents the old-fashioned way or with the then horribly limited word processing systems.
Koplow and Moros wrote a manual for an idealized word processor without any consideration of implementation. When they were done with it, they pitched the idea to Dr. Wang, who gave it an enthusiastic approval, charging them with actually building it.[17]
Rather than the mainframe-style of centralized computer, they decided to use an Intel 8080 microprocessor in each terminal to make it responsive, and then to network the heads to a central file server.
The Wang WPS (Word Processing System) was introduced to the world in June, 1976, and was an immediate success. Within three years, Wang's sales tripled, mostly on the basis of the new WPS sales and the synergistic effect it had in selling other Wang hardware.[18]
Wang OIS (link)
The single WPS system quickly evolved into a more general family in 1977, the Wang Office Information System, or OIS for short. OIS had even more capable word processing capabilities, a general programming language for more technically adept customers, and greater networking options. Each component in the system, terminal, file server, printer, contained memory and a microprocessor, and were well thought out to allow a customer to configure a system anywhere from an individual word processing station to a fully networked office.[19]
Wang VS (link)
Wang introduced the VS family in 1977 as well. Based on the IBM 360 instruction set with some small changes, Wang developed their own OS from scratch. This family was aimed at the mainframe corporate customer. However, Wang made sure they had a version of their word processing software which also ran on the VS machines. Although Wang was at first afraid of the VS and OIS systems cannibalizing sales from the other, in fact it seemed to act in synergy.[20]
Footnotes (link)
- ⇑ Riding the Runaway Horse, , p. 42
- ⇑ Lessons, An Autobiography, , p. 120
- ⇑ Ibid.
- ⇑ Ibid., p. 121
- ⇑ Riding the Runaway Horse, , p. 44
- ⇑ Lessons, An Autobiography, , p. 173
- ⇑ Riding the Runaway Horse, , p. 59
- ⇑ Lessons, An Autobiography, , p. 174
- ⇑ Annual Report, 1976, , p. 22
- ⇑ Lessons, An Autobiography, , pp. 175-176
- ⇑ Ibid.
- ⇑ Photocomposition Input Option for the Wang System 1200, , 12/74
- ⇑ Annual Report, 1976, , p. 22
- ⇑ Wang Printout, Vol. V, No. 3, , p. 8
- ⇑ Annual Report, 1976, , p. 20
- ⇑ Microchip: An Idea, its Genesis, and the Revolution it Created, , p. 186
- ⇑ Riding the Runaway Horse, , pp. 64-70
- ⇑ Ibid., pp. 71-75
- ⇑ Lessons, An Autobiography, , pp. 205-206
- ⇑ Ibid., p. 207