History can be written in many ways. This includes the history of technology and computer science. One way of looking at things relates to often the context in which something appeared. Let's take the "western" type of keyboard layout which many of us are used to: the qwerty keyboard.1
An argument of why we have this might not be as interesting as the question of why it became som popular and wide spread. The question of why can be construed in hindsight (even contrafactual), and might not be the actual reason (or in case: reasons) for why it became to be so popular. The reason, it could be argued, the keyboard layout was due to technical reasons, as the (maybe overlapping) typebars should not collide to often (as the often used keys are spread out). Those of us who have typed on an ordinary typewriter may agree to this, and draw the immediate conclusion that this must have caused the layout to be the still dominant way of manufacturing keyboards today. This historical "hunt" for such causal relationships may or may not be the actual case, naturally. But, perhaps, rather than internal arguments, we could look for emergence of streams of technological innovation where some find their way through the changing landscape and derive larger and larger streams, but where others stops and go no futher. To keep with the analogy, success (or failure) might not have to do with the strength of the stream itself from the start, but may very well have to do with the form of the landscape, hills and valleys, soft and hard ground, and add to this the constant change through time (which deviates from the impervious landscape).
We can thus speculate in the historical success of "qwerty" as the result of a lot of 'things' influencing its path. One of these influences can be the technological reason (overlapping typebars) mentioned above. But it not even might be the cause, if such a thing actually exists. Other concerns might have been sale prices, availability of sources or ready made items, unfortunate or fortunate events, unforseen events, etc. etc. But context may clarify circumstance.
Image from patent: https://patents.google.com/patent/USD224415S/en. 2
In 1970 Computer Terminal Corporation (CTC) annonced the terminal/computer Datapoint 2200.3 Later, in 1971, the company shipped the product to customers. The rather recent video terminals4 had some inheritance from the old teletype.5 But as many video terminals relied on replacing the interface to the computer from teletype, papertape and the like, the new Datapoint 2200 introduced more of a proper computer inside the case, besides being a terminal emulating Model 33. This was innovation. Therefore there are some claims as it was the start of the computer that was personal, or private to the operator, i.e. a "PC".6
There was several periferals that could be connected such as removable cartridge hard disk drive, modems, priters, serial and parallel interface, but also, later on, it was first with an 8" disk drive. But also there was software which could independently run, not only emulate different terminal protocols. It had its own processor, its own CPU. The idea from start was that a CPU could be designed such that it would be possible to have the main functions in one chip (to reduce problems with heat). CTC designed this chip called 1201, and went to Intel7 for planning ultimately in production. Also Texas Instrument was asked to compete. However, after some time, CTC went ahead and made the CPU by TTL-logic in hardware instead. They were not satified with what was delivered to them. TTL-logic was thus the hardware base when Datapoint 2200 first went into production. At the time CTC dropped the idea of a CPU on a chip, and transferred the design to Intel.8 This happend almost concurrently with the design of Intel 4004 for Busicom and their calculator, released in 1971.9 At this time memory was the really expensive part in machines, and therefore decisions on design and production had a different focus than a decade later on. A CPU (or part of a CPU) on a chip would however reduce costs as well as the heat issues, even though it were some issues with speed in the first of designs.
The Datapoint 2200 was very popular, and had a relatively long life. It also changed its internals over time, improving e.g. memory from serial shift register to RAM, or adding a hard drive. On the other hand it could only reach a limited set of customers, as it was leased and not sold at the start, exactly as IBM's products were at the time from which the marketing model was copied.
A renewed interest in the 1201 design came up as Seiko found interest for a scientific calculator. An improved design turned into Intel 8008, a microprocessor that could arguably claim to have started the microprocessor/ microcomputer revolution. But also the lineage to its processors for ordinary PCs of today in 'x86'.10 This is one of the strains from the stream. The fact that the core of instruction set, or the use of little endian kept making progress for decades to come, was the demarkation of a strong hold. A physical habit, one might call it. Hard to overturn.
Advertisment from Datapoint (recently renamed from CTC) in Computerworld, December 1972, p.12.
Seven-S. Image from Peter Häll - https://digitaltmuseum.se/021026362412/dator, CC BY 4.0, https://commons.wikimedia.org/w/index.php?curid=80713670
The man behind the technical part of ABC80, Lars Karlsson, started in 1971 his company 'Dataindustrier AB' (approximately translated as "Computer Industries ltd.").11 The core products was for a long time based on the bus "Data Board 4680". The numbers in the name came from the intended support of Intel 4004, Motorola 6800 (later also MOS 6502) and Intel 8080 (later also Zilog Z80). Besides a versatile board computer driving this bus (4/8 bit processors), cards could be added in the style of the later popular S-100 bus12 to build custom computers suitable for the intended audience of industrial applications.
In 1974 a memorandum describes ideas for a new computer, the "Seven-S". At the end of the memorandum, the main competition is mentioned:13
The main competitor is Datapoint 2200 made by Computer Terminal Corporation, USA. This unit has been on the market for about four years and some 6000 units have been sold, of which 300 in Sweden.
Datapoint 2200 has been a success and has open a market for a new generation of decentralized computers. Many new brands will follow but Seven-S will take an economical and technological lead, and based on aggressive marketing, will reach most of the potential markets inside and outside of Sweden.
Seven-S will be the new weapon to fight the computer giants like IBM, Honeywell-Bull, CDC etc.
Maybe a bit ironic is the cirumstance that DIAB Data AB (as the company later was called), was bought in 1991 by Bull, later in 1994 denationalized in France.14
The functions of the computer is described such as it could be used for word processing, data collection, as a terminal, remote batching, stock-keeping, inventory, personnel, salaries, etc. suitable for then current office needs. A remark is also that smaller companies could use this computer, as well a larger (through the terminal function). Also other applications are possible, programmable calculator systems, process control, and industrial programmable controller system are mentioned. In total this made up the seven bullet points, the seven application areas, to which Seven-S was intended, hence its name.
In 1975 eventually Dataindustrier AB together with Innovation Tomas Nilsson AB, formed a new company, a joint-venture, Data Future AB for the sole purpose of selling Seven-S. It was promising at first, they acquired some needed investment capital from the state, they won a reputable prize for innovation, and also a larger customer (Kommundata) was ready for 500 terminals, or even more in the future.
Hardware isn't mentioned much from what can be found in sources, but was build around the Z80, with 16k byte of dynamic RAM, initially at least from the outset according to an article.15 From the picture displayed above, the caption says that the configuration here is a Zilog Z80 processor, it has a 15" screen (25x18 characters), and 64k of RAM. The Seven-S was manufactured by Stansaab starting 1977, which later fusioned with Datasaab. Production was relatively slow.
Then something happend. The control of the sole customer and the product was transferred to Datasaab.16 Datasaab cancelled the project in 1978/1979, and that was the end of that. A total of hundred Seven-S were produced.
Ignoring interesting futher politics here involving larger companies and competition, the ideas and solutions from Seven-S was brought in to the project of ABC80. The pace in which the project of ABC80 proceeded could only come through the previous experience of its predecessor Seven-S. A small strain survived.
- A short history on Datapoint 2200: https://www.old-computers.com/museum/computer.asp?c=596
- More on Datapoint 2200: https://web.archive.org/web/20080819031221/http://www.computerworld.com/action/article.do?command=printArticleBasic&articleId=9111341
- An archive about Seven-S: https://www.df.lth.se/~triad/diab/archive/Seven%20S/, or https://dflund.se/~triad/diab/archive/Seven%20S/
Footnotes
-
Seems to show Datapoint 1100 which had one cassette interface, rather than two as in Datapoint 2200, as only one lid on the top is visible. ↩
-
https://en.wikipedia.org/wiki/Datapoint_2200 and https://en.wikipedia.org/wiki/Datapoint ↩
-
There is often, to my mind, a futile debate on who came first (of anything). Depending on what criteria you use, different computers come to mind. One compeditor here should probably be Kenbak-1: https://en.wikipedia.org/wiki/Kenbak-1 sold in 1971. The French have their Micral: https://en.wikipedia.org/wiki/Micral in 1973. The Swedes (of course) also claimed to be first: https://dflund.se/~triad/diab/archive/Tidningsartiklar/1986%20Datornytt%2010%20om%20Transdata.jpg. For our story here, the interesting part is that Lars Karlsson (together with others) used the Intel 8008 in a computer intended for education in e.g. programming: Transdata 7261. It thus had a microprocessor and could mark the start of the microcomputer. It was introduced at a computer fair, one kilometer from where I grew up, Älvsjömässan in 1972. Besides the Transdata in 1972, there was a Univac 1972 and a Q1 from Belgium. All of them had the 8008. All of them were born in 1972, including the processor. The Q1 can be seen in the video: https://www.youtube.com/watch?v=dB3V_Q9wQ-M&t=6s. The Univac 1982 can be seen at: https://www.youtube.com/watch?v=9KojS1ezQIY. And the Transdata have no video, but some presentation at: https://sv.wikipedia.org/wiki/Dataindustrier_AB (in Swedish), and e.g. https://dflund.se/~triad/diab/archive/Transdata/Transintro%20Transdata%207260.jpg (also in Swedish). ↩
-
Or e.g. previous more accutrate influencial DEC UNIBUS (https://en.wikipedia.org/wiki/Unibus). ↩
-
See https://www.df.lth.se/~triad/diab/archive/Seven%20S/1974%20SevenS%20konceptstudie.pdf (archived by Linus Walleij). ↩
-
Dagens Industri, 14 Feb. 1978 (?) ↩
