Shown here on the cover of Electrical Experimenter a hundred years ago this month, May 1918, is Hugo Gernsback‘s vision of Skype, or what he called “telephot, an apparatus attachable to our present telephone system so that when we speak to our distant friend, we may see his likeness not only as an immovable picture, but we will see his image exactly as we see our own image when looking into a mirror. In other words, the apparatus must faithfully follow every movement of our distant friend whether he is only five blocks away or one thousand miles. That such an invention is urgently required is needless to say. Everybody would wish to have such an instrument, and it is safe to say that such a device would revolutionize our present mode of living, just as much as the telephone revolutionized our former standard of living.”
Gernsback reported that inventors were working on the problem, but the main catch was what we would today call the bandwidth, since it had to be “possible to attach it to the present-day telephone lines,” which to him was a single wire and a ground return wire. “In most of the schemes offered by inventors heretofore, a plurality of wires was necessary; in some cases several thousand pairs of wires. No matter how well such an instrument might work, this alone would doom it to certain failure.”
Seventy years ago this month, the cover of the May 1948 issue of Radio Craft showed the latest in electronic computing equipment, this 12,000 tube electronic brain, the IBM Selective Sequence Electronic Calculator.
Along with those 12,000 tubes, the computer contained 21,400 relays and 40,000 pluggable connections. Power consumption was about 180 kilowatts, and the machine had an extensive fire suppression system built in. The accompanying article began with an example of computing the position of the moon, given a starting point of December 31, 1899. This task would take a human about three weeks, but the IBM was able to crank out the calculations in a mere seven minutes.
The computer covered the walls of a specially designed room measuring 40.6 by 86.6 by 14 feet. In a minute, it was capable of 3500 additions or subtractions of 19 digit numbers, 50 multiplications or 30 divisions of 14 digit numbers. Logarithms and trigonometric functions were stored on magnetic tape, and when one of these functions was required, the required value was looked up from the table prerecorded on the tape. Punched cards or magnetic tape were used for storage and input and output.
Memory during the arithmetic functions was handled by trigger circuits using ordinary 12SN7 dual triode tubes.
According to IBM President Thomas J. Watson, “this machine will assist the scientist in institutions of learning, in government, and in industry, to explore the consequences of man’s thought to the outermost reaches of time, space, and physical conditions.” But the machine “cannot, and never will, replace the human scientist as its purpose is only to follow his commands. If the incorrect instructions are given to it, it will follow them. The scientist is supreme over the great electronic calculator, which is his child to assist him in exploring the ever-profound depths of science.”
Forty years ago this month, this couple is enjoying a friendly game of Pong, as shown on the cover of the January-February 1977 issue of Elementary Electronics.
As detailed in the accompanying article, they put the game together themselves, thanks to the Interfab Pong-IV video game kit. The unit contained 43 integrated circuits, and came in three forms. For the purist kit builders, the set came with all of the parts, and the builder had to populate the circuit board and solder them. To save a considerable amount of labor, it was also available with the board pre-populated with parts, held in place by a plastic blister pack. The builder then merely had to solder the multitude of small connections and then remove the plastic. Finally, it was available in semi-kit form, with the circuit board already soldered, and only minimal mechanical assembly required.
The kit was originally marketed with a built-in UHF transmitter to hook directly to the TV. However, the FCC cracked down and required type acceptance, which wasn’t economically viable. Therefore, the PC board was all ready to go, and the manufacturer provided a parts list and instructions to install the transmitter, using a 2N5770 transistor and a few other parts. Other options were to separately purchase a UHF transmitter, for a cost of about $8.50, or tap right into the TV’s video amplifier (this was before the days of most TV’s having a video input jack).
The kit was offered by the Interfab Corp., of Laguna Niguel, California. The completed kit sold for $89.50, with the less assembled versions being about $10 less.
Thirty-five years ago this month, the November-December 1980 issue of Elementary Electronics carried this ad for the first personal computer for under $200, the Sinclair ZX80.
The computer was most famous for the membrane keyboard. It was impossible to touch type using this keyboard. One popular modification for later models was the addition of an external keyboard. It contained 1K of memory, some of which was used for the display. This could be expanded up to 16K with an external memory module. As the name implies, it used a Z80 CPU chip, and the ROM came pre-loaded with the BASIC programming language. The output was a monochrome TV signal. Since the computer used the bare minimum of hardware, the generation of the video was handled by software. Therefore, while the computer was actually computing, the screen went blank, which one reviewer noted had the small advantage of letting you know that the computer was actually working.
Program was handled by an external cassette recorder.
In addition to the ad shown above, this issue of Elementary Electronics also carried a review of the computer. It noted that the version of BASIC included was uncommonly extensive and flexible for such a low-cost machine. The review concluded that while the ZX80 was not a substitute for a full-size computer, it was a low-cost way to get into personal computing and learn programming. Indeed, another review pointed out that the cost of the computer was less than a college course in BASIC programming.
A couple of years later, a more familiar version of the computer came on the market as the Timex Sinclair TS1000. It came on the U.S. market in July 1982. With a retail price of $99.95, it was billed as the first computer under $100. The price soon dropped to $49.95. Competitor Commodore’s VIC-20 was somewhat comparable, but had a full-sized keyboard. When Commodore announced that it would offer a $100 credit for the trade in of $100 on any competing computer, many TS1000’s were sold for the sole purpose of trading in on a Commodore 64.
My first computer was another competitor, the TRS-80 MC-10, which originally came on the market in 1983 with a price tag of $119.95. While designed to compete with the Timex Sinclair and having similar capacities, it did have a number of advantages. While it did not have a full-size keyboard, it did have actual keys, rather than the membrane keyboard of the Timex Sinclair. Also, it had full color and even had some rudimentary graphic capabilities. I bought mine for about $49.95.
The calculator complete with the new added functions.
In the early days of electronic calculators, one common IC was the MM5737, which supported four functions and eight digits. Its big brother, the MM5738, was slightly more advanced, since it included a single memory, a constant function, a percent function, and a battery-saving feature that would turn off the display after about 60 seconds.
What gave early hardware hackers (long before the term was invented) something to do was the fact that many calculator manufacturers stocked only the MM5738, even though some of their calculators didn’t make use of the extra functions. The cost of the chips was about the same; they simply didn’t wire in all of the functions on the less expensive “four banger” calculators.
Someone at Popular Electronics noticed this, and figured out how to add the hidden features to the less expensive models, which was revealed in the September 1975 issue. The first step was to determine which chip was contained inside, and this could be done from the keyboard, without even opening up the case. This was because the MM5738 had the ability to peform repeated squares. From the keyboard, you simply had to enter 3, x, =, =. If the display said 81, then there was an MM5738 inside. If the display said 9, then the calculator used the more basic MM5737, and no modification was possible.
The author acknowledged that there would be no economy in trying to find a keypad with the added buttons. Instead, he proposed mounting four pushbuttons on the side of the case, to serve as the MS (memory store), MR (memory recall), K (constant), and % (percent) keys. He performed the modification on a Novus model 850.
I earlier wrote about the Novus model 650, an even more bare-bones model that lacked a decimal point. I suspect it used the same chip, and a similar procedure could have added it. The 650 had a retail price of $19.95 when it came out in 1974, but was down to $8.88 by December 1975.
Armed with the information in this article, owners of one of these very basic calculators could beat the system and save a couple of dollars by upgrading to a more advanced model by themselves.
40 years ago, careful observers were probably anticipated that something like the Internet would sometime come to pass. For those early adopters, Radio-Electronics magazine carried plans for a terminal, which the magazine called the TV Typewriter II, since an earlier version had appeared in the magazine in 1973. Some of the chips used in the 1973 design, however, turned out to be unobtainium, and the 1974 model featured use of commercially available 74-series TTL chips. The plans included a keyboard (hence “typewriter”), and called for connection to a black and white TV for the display (hence “TV-typewriter”). The plans for the 1974 model, shown here, are found in the February, March, and April, 1975, issues of the magazine.
The terminal had an RS-232 interface, and could communicate at 110, 220, 440, or 880 baud, or if a different crystal was used, 150, 300, 600, or 1200 baud. The terminal had two potential applications. First, the article noted, would be “data communications with computers. Combined with a keyboard, we have one of the fastest and most efficient means for an individual to communicate with a machine.” One example of a computer with which it could communicate would be the fledgling home computer described in the magazine’s July, 1974, issue.
Another, possibly more common, use for the TV-typewriter was also given: “Of course, if you don’t have or don’t want your own machine, you can always tie into a full size time-shared system, assuming you have access to one.”
This statement shows how the TV-typewriter foreshadowed the Internet. My first experience with computing involved one of those time-shared systems. In my case, I accessed the Minneapolis Public Schools’ timesharing system with the use of a 110-baud teletype machine, like the one shown here. To use this terminal, we dialed the phone number of the timeshare system computer, placed the telephone receiver into the acoustic coupler modem, and after logging in we were able to do things such as write short basic programs. Later on, other systems such as the Minnesota Educational Computing Consortium‘s MECC-MERITTS system, which also included a rudimentary chat function.
Of course, accessing such a system required having physical access to a terminal, which nobody other than the school owned in those days. Something like the TV-Typewriter would have been an incredible luxury for those of us interested in the proto-internet offered by the timesharing systems.
Many of us back in those days figured out that it was just a matter of time before something like the Internet came into being. However, most of us were wrong in one regard. We believed that online computing would follow the model of those timeshare systems: There would be a big computer somewhere doing all of the computing, and users would access it using a dumb terminal like a teletype machine or the TV-Typewriter.
In some respects, our predictions were borne out a few years later when CompuServe came along in the late 1970’s and early 1980’s. It was way out of my price range, at about $5 per hour (a rate at which it would cost about $20 to download the entire contents of a daily newspaper). But since I worked for Radio Shack, which was one of its primary sellers, I did have limited access and was able to use it to read the handful of newspapers online, get aviation weather, and even access the early chat server. There was even something called “Email,” a registered trademark of CompuServe. It was still out of my price range, and I never even sold a single subscription, but I realized that it was the future of computing.
Except I was still wrong on one detail. I still assumed that online content would be delivered the way that CompuServe was doing it: Individual users would use a dumb terminal to call a big computer somewhere that delivered the content. The 1981 video below is a newscast about the early days of CompuServe. It also makes clear that this is the model: Once a day, the newspaper in San Francisco programs into a computer in Columbus, Ohio, the contents of the daily paper.
Most of us believed that was how it would work, and we were wrong. We probably could have figured this out, because to access that computer in Columbus Ohio, we weren’t exactly using a dumb terminal. We were using a computer (at that point, the TRS-80 Model 100). By today’s standards, the computer didn’t have much computing power. It had 8 kB memory and ran at 2.4 MHz. But that computing power was doing only one thing: Emulating a dumb terminal. The computing power wasn’t being used, other than to duplicate the functions of the 1975 TV-Typewriter.
By the early 1980’s, the first computer bulletin board systems (BBS’s) were starting to appear. They esentailly followed the same model as the early timesharing systems and CompuServe. The only difference was that they were much smaller systems, generally run by hobbyists. But the end user still called up the “big computer” with a dumb terminal. The real origin of the internet is best seen in a development that took place a few years later, when BBS systems began to be linked up as FidoNet. FidoNet was developed starting in 1984. Users would still call in to a single local BBS. But at night, when the phone rates were low, those BBS’s would communicate with one another and transfer mail, discussion posts, and files. While the end user would still interface only with one local “big” computer, the information he or she was reading actually originated on some other computer. In other words, the information was no longer centralized in Columbus, Ohio. The absence of one computer from the larger network would hardly be noticed. For home users, the Internet really began in 1984 when FidoNet allowed them to access information globally from a dumb terminal. Most users of these BBS systems were using computers with some computing ability. I doubt if too many TV-Typewriters were being used in 1984. But the TV-Typewriter, assuming that someone got one working in 1975, would have been perfectly adequate to the task.
The TV-Typewriter by itself was still missing one important component. It did not include the modem. To use it with a timesharing system (or a few years later, with CompuServe or FidoNet), the builder would also need to construct a modem to get those 110 baud signals into the phone line.
For those who built the TV-Typewriter, they probably had some inkling that there would be an Internet someday. They had the details wrong. They probably thought that there would be a big computer in Columbus Ohio to which they would connect. But they probably had some of the basic ideas.
The home computer is 40 years old. The one that appeared in January 1975 issue of Popular Electronics. Used copies of this issue typically fetch about a hundred dollars on eBay, but fortunately, a full scan of the issue is available at AmericanRadioHistory.com. The January issue carried a summary of the computer and some of the construction details. The February issue included an introduction to programming it.
The January issue carried an editorial announcing that the home computer was here. It correctly noted, “for many years, we’ve been reading and hearing about how computers will one day be a household item. Therefore, we’re especially proud to present in this issue the first commercial type of minicomputer project ever published that’s priced within reach of many households–the Altair 8800, with an under-$400 complete kit cost, including cabinet.”
The construction article billed the computer as the “Popular Electronics/MITS Altair 8800.” It was built around an Intel 8080 CPU chip, which could handle up to 78 instructions. The construction article did contain a parts list, but not full PC board templates. Those were available by mail, but it’s likely that most builders took advantage of the computer’s being available in kit form for $397, or fully assembled for $398 from MITS, Inc.
The basic computer came with 256 words of memory, with up to 65,000 being available through add-ons. The parts list called for a 2 MHz crystal, indicating the processor’s speed.
The January article suggested some possible applications for the computer, such as use as a programmable scientific calculator, machine controller, or automatic drafting machine. The February issue included the basics of programming the computer, along with a sample program to add the contents of two of the registers and store them in a third.
Programming was accomplished from the spring-loaded switches on the front panel. Another possibility for programming the computer was to use a computer terminal, and the article suggests the design that had appeared in the December 1974 issue, shown here. It’s probably not the mental image that would come to mind upon hearing the phrase “computer terminal.” It’s simply a method of sending an octal code to the computer, and receiving one back.
It was a year or two after the computer first appeared that I first saw one. When I saw it, the builder hadn’t really come up with anything for it to do. He was working on interfacing it with a teletype machine, and at that point, all he could make it do was have it output the character associated with a particular ASCII code. In other words, he entered a number using the front panel switches, and it printed out the corresponding letter. I remember not being too impressed, but I guess I did realize that at some point, I might have a computer in my house.
Shown here on the cover of Electrical Experimenter a hundred years ago this month, May 1918, is Hugo Gernsback‘s vision of Skype, or what he called “telephot, an apparatus attachable to our present telephone system so that when we speak to our distant friend, we may see his likeness not only as an immovable picture, but we will see his image exactly as we see our own image when looking into a mirror. In other words, the apparatus must faithfully follow every movement of our distant friend whether he is only five blocks away or one thousand miles. That such an invention is urgently required is needless to say. Everybody would wish to have such an instrument, and it is safe to say that such a device would revolutionize our present mode of living, just as much as the telephone revolutionized our former standard of living.”
