Is there a reason why computers of the 1920s-1980s look the way they do? [closed]

Question

Good day all.

I'm currently penning a document for my worldbuilding project that deals with the aesthetic (or rather design language) of the world's current level of technology. For reasons indistinguishable from mere whim, I found myself very interested in the computers of the 1980s, and extrapolations from their own design language that constitutes an entire aesthetic called 'cassette futurism'.

I've done a fair bit of research on it (constituting mostly scrolling images and taking notes) and it led me to look at various media ranging from Evangelion, to Akira (still need to watch that one), to Alien and other science fiction classics of its time, and somewhat back to our world in the form of the Apollo moon mission's; both the now deprecated mission control of the time and the cockpit of the lunar lander. It was about when I stumbled upon the German 'technikum29' museum of old computing that I finally pondered, why really do they look like that?

It's easy to grasp why technology in our age looks the way it does; an endless push for more sleekness, more lightness, making hardware less obtrusive and software somehow more so. Am I correct in doubting that years ago, the doctrine was that computers had to be large, bulky, obtrusive, cubical, built of plastic and packed with buttons and sliders and arrays of blinking lights? Is this design merely a product of that time? Or is there something more?

In my own canon, my justification is that design doctrines of today - sleekness, lightness, touchscreens and voice controls, making things smaller and smaller and smaller - never really caught on due to the time period where computers became more widespread being one of multi-stellar warfare. The protagonist faction is defined by a cosmic nomadism that sees them crossing stellar distances via a nigh-genetic supernatural power and various flavors of FTL-capable vessel.

Their technology was, as such, a product of this, built to be robust, durable, easily repairable. The first major use cases for computers were in vessels built for interstellar dogfights, and in time it was discovered how much more capable a pilot was when provided opportunities to develop muscle memory and be rewarded with tactile feedback. Analogue controls, dated controls came to rule the day as a result, and they were disinclined towards changing from this even when the warfare ceased. Perhaps an unrealistic human reaction but I digress. The earliest civilizations lived in harsh conditions, and thus their technology came to reflect that.

I am curious if anything along those lines drove how our own technology decades (the 1980s to be precise) ago came to take on this distinctive look. Thanks in advance.

Answer 1

Reliability & Failure Awareness

I did a job once that involved taking a tour of the control room of a nuclear power plant, and much to my suprise, there was virtually no computers in the whole thing. They explained that they have no plans in the future to ever switch to a computerized system because so many other power plants that did proved to be far less safe and reliable, and they represent major vulnerabilities in the country's infrastructure.

When you are dealing with a system where failure is not an option, computers can do more harm than good. When you place a computer between a sensor and that sensor's output, that computer becomes an unnecessary point of failure. If the computer gets hacked, then you could be forced to shut down a reactor for days or even weeks to diagnose a problem that is not actually there. Or worse, a glitch could cause your system to fail to alert you that your reactor is about to melt down. While a sensor failure is mostly a non-issue for most practical purposes, when millions of lives or a multi-billion dollar starship is at stake, they can not be allowed even once.

Part of what makes those old school control rooms look the way they do is that each system gets 4 redundant sensors for every single function. If one sensor goes out, the remaining sensors can continue to confirm the proper functionality of the system, but more importantly, if one is just mis-reading, the other 3 can catch the miscalibration where a single sensor system would appear to still be working while giving the wrong information. Because each sensor is on a seperate circuit connecting to each output, there is no chance that a single point of failure will cause all 4 to fail at once, but when you put a computer, network of computers, or even un-networked computers running identical software between them, it's much easier for something to go wrong to make all 4 simultaneously give a false result.

The War impacted Culture to always demand verifiable operability

In our culture, the first generation of PC users were children who (as children do) believed that nothing can or will ever go wrong until proven otherwise. They did not care if a PC could prove that it was working right because the idea that it might not work right was far from thier minds. So, all those lights and durability features on early mainframe computers just seemed unnecessary. They did not market well; so, they were removed in favor of the "smaller is better" mentality.

However, in your setting, the first generation of PC users were not children. They were the millions of paranoid veterans returned from war having come back from seeing thier buddies turned into stardust over cyber warfare and system failures. So, they came back with a military devotion to making sure thier tech is never compromised, and that software must never be trusted 100%. If thier PC does not have 4 LEDs directly tied into the power supply to make sure it is functioning to spec, they don't trust the computer at all. If there are not physical gauges on the computer reporting CPU, Memory, and Disk loads, then they don't trust that the computer is actually running as it says it is. After all, Windows Task Manager is just software, and software can lie to you.

If you consider that each key system should have 4 sensors and a basic PC has many possible points of failure, it's easy to imagine this mentality requiring the front of a PC to be covered in status lights, even if the other hardware specs are exactly identical to modern PCs. Basically, all those blinking lights become a marketing tool to make the average consumer feel more confident about buying your product. The Reliability mentality also extends to the physical shape of a PC. People want something that LOOKS like it won't fail; so, most consumers will pick the thing that looks like an armored safe over the thing that will save them a few square inches of table space, even if the case is just cheap plastic and mostly empty space.

Even a generation or 2 out, the sleek fad will still not have caught on because your world has one thing that ours did not. Parents well equipped to teach thier kids about computer safety and reliability. Computer aesthetics will still change as fashion does what it does, but the goal of computer design will remain to make something that looks trustworthy, not something that look cute.

Also: slide cell phones need to beat touchscreen phones to market

sleekness, lightness, touchscreens and voice controls, making things smaller and smaller and smaller - never really caught on ...

The thing about smaller and simpler with cellphones is that there is a very measurable and practical reason that smaller is better when designing a device meant to be carried with you at all times. There is no getting around the idea of a pocket computer being a good thing, but you can make them evolve differently.

The first touchscreen device was invented in 1970, but it was cell-phone companies that made them popular. Touchscreens were a total flop the first few times people tried introducing them because they are by nature slower and less ergonomic to type with than a keyboard and harder to achieve precise operations with than a mouse. But the demand to make a phone with more screen space was high enough to justify continued research and development. Touch screen phone were already around for almost 10 years before you start to see slide screen phones that could practically fill the same need. Most users preferred the slide screens when they came out, but Apple, which produced the most powerful cell phones, saw them as redundant and decided not to offer any phones with this feature. The end result was that people looking for the power to use thier phone like an actual computer had to use Touchscreen exclusive phones and people going for cheap did not want to spend extra on a slide screen, so the slide screen became obsolete almost as soon as it was introduced because it was the sort of unwanted middle option.

But, if the slide screen came first, then the first few touchscreen flops would have convinced the tech industry that touchscreens are too much of a risk to invest much further development into.

Answer 2

The 19-inch rack was established as a standard by AT&T around 1922 in order to reduce the space required for repeater and termination equipment in telephone exchanges. This was a useful standard for packing electrical equipment into towers with a common power supply. I cannot identify any 19-inch racks in pictures of ENIAC, but they do have things sorted into similar-looking units. Later, IBM, DEC, Honeywell, Burroughs, Univac, and other mainframe makers used the same units. They did like having lights on the front, but I feel the function (are all the boards getting power) came first, and the front plate was the last bit to go in. If there was an aesthetic, it was to get lots of identical-looking units even when the contents may be different. That somehow meant 'power'.

When I started working on motion pictures, servers, tape machines, video players, and all kinds of other stuff were loaded into these 19-inch racks giving walls of blinking lights very like the Millennium Falcon. I don't know, but it is not hard to see how one might take inspiration from the other.

Answer 3

CRT Tubes

This isn't really a question that has a right answer, but my personal take is that the technology aesthetic that you see in "cassette futurism" was, in a large part driven by the technological limitations of the time, in particular, the volume requirements of CRT monitors.

Below is an image of a cutaway CRT television [source]:

Here, the actual tube is extremely "deep" compared to the flatscreen monitors we have today: the ratio is almost 1:1 in a screen-size-to-tube-depth way, and this is a major part of what lent a lot of the "bulkyness" and "boxyness" to old electronics.

In addition to bulky CRT modules, electronics were just bigger back then. We couldn't manufacture chips at small scales, which mean that many large and bulky discrete components had to be housed within the casings to perform functions that we can put onto a chip the size of a grain of rice today.

Sidenote on "robustness" and "reliability"

You seem to be conflating "analog controls" with intrinsic factors like robustness and reliability, which isn't really true. For example, many modern aircraft are transitioning away from manual switches, buttons, toggles, and dials wherever they can, instead opting for digital multi-function displays (MFDs). This is for a variety of factors, but "sleekness" isn't really a driver.

Rather, it's that digital displays are not only more capable, but in many cases, also more reliable than mechanical dials or switches which, due to their mechanical nature, are prone to wear and eventual failure. In fact, if you are building an aircraft yourself (like from a kit), basically all modern guides advise builders to go with digital systems because not only are they lighter weight, easier to use, etc, but they are also more reliable than a system with many moving parts.

Answer 4

Am I correct in doubting that years ago, the doctrine was that computers had to be large, bulky, obtrusive, cubical, built of plastic and packed with buttons and sliders and arrays of blinking lights? Is this design merely a product of that time? Or is there something more?

Well, you are correct in doubting that they had to be like that. But, the computer design was not just the product of that time. The problem here is threefold:

Maintenance

First computers' components were vacuum lamps, which had to maintain characteristics for a long period to allow that computer to actually compute something that isn't line noise. And those lamps were a pretty new invention by that time; there was a number of longevity problems with them that yet had to be solved when the first computers were designed. Thus, their design had to allow easy access to its insides in order to diagnose and repair or replace a faulted lamp or whatever other element that failed or went outside the required parameters. After all, any computer is an analog circuit, it's just how we interpret signals (and make it interpret its own signals) that makes it digital; also there were and probably still are specialized analog computers, designed to solve a single analog task or a range of them, determined by wiring and correct use of resistors and such, that still do their job with enough precision and speed to not transition to purely digital computation. So, early computers could fail a multitude of ways, and a lot of time was spent calibrating the timings on their processing circuits to produce the effectively correct signals at correct times on every element, so it would transition to the correct state and itself produce the correct signal to subsequent elements. In order to make maintenance easier, some components had state lamps wired into the circuit (which were accounted for at design time), that could be used by an electronician to quickly find out if this particular component works correctly.

Output

Also there were no display screens back then, thus the computation results had to be displayed somehow else. Some of the computers used a set of lamps as a means to output a value, in form of them being lit or not, that if read in binary would compose the result, and such sets of lamps were actually called "registers". These, based on the program being run, could also store some intermittent values until the program completed, then once they stop blinking, the final value could be read from them and recorded as a result. Sometimes the operators made more runs of the same program, with the same input data if expected (early the program's data was bundled with it on the same media), in order to verify if the computer didn't break in the middle of execution, an experienced operator could detect that there was an error by verifying if the same light pattern happened during each execution.

If the expected result was longer than the set of registers, the output was sent to a printer, which itself wasn't a standalone device, but instead a wired module that could also malfunction and required diagnostics frequently. This device had to be both close enough to the central processing unit (not CPU mind you, it wasn't a chip, but a whole box filled with wires and lamps and stuff) to not lose or distort the signal, and close enough to the operator to not have to run over to the far end of the computer room just to read another 0xDEADBEEF in the printed line.

These two aspects combine into the third which happens to be a tad more than the sum: Functionality. The designers had to place all the elements of the computer somehow so that they are easily accessible in case of failure, and that the computer's input and output subsystem(s) are located conveniently to the operator. And also to not let a fumbling operator break the computer by leaning on it. So the design became boxed, lamped, wired and switched (tapes came pretty late and replaced switches). In fact, the science of design didn't come up until pretty late into the human history, at first people required functional equipment, then some of it could be made aesthetically pleasing (which increases cost), and only when it came to mass production, the people started thinking about design as a separate quality from durability and convenience. And it took the humanity several decades to make design transition from art to science; and I dare to say this transition isn't finished yet.

And finally, the early computers were required by armies, where design is not valued at all, unless it somehow disturbs normal operation, at which point the design is dropped in favor of making the dam thing work. So that day's design was not "just" a product of that time, it just served purposes not deemed necessary by today's users.

Answer 5

Most computers, then and now, were in metal boxes. The dimensions of the box were often determined by an industry standard. So, while AT and AT-tower cases became standard for home PCs in the late ’80s, so that builders could add components from many vendors, IBM also made a rack-mounted industrial-control version of the AT.

Since a computer needed a keyboard and a display to be useful, many models had the monitor and keyboard built into a single unit. This originated with the terminals that could plug into mainframes, which originally had no processing power of their own, such as the IBM 2260 of 1964 and IBM 3270 of 1971. These sent whatever the user typed to the central computer, and displayed whatever it sent back on its screen. The DEC VT100, which did not run programs on its own but had the ability to interpret complex formatting, became an industry standard. Their keyboards were like the mechanical typewriters that office workers were trained to use. It also became the form factor of the earliest business computers, such as the Wang 2200 dedicated word-processor and the Hewlett-Packard 9100 desktop calculator. By the time computers were cheap enough for individual office workers to get one, they were small enough to fit on top of a desk. This also became a popular form factor for school computers, starting with the Commodore Pet, because it was harder to remove and steal a single component.

Home microcomputers, on the other hand, were originally designed to be cheap and plug into a TV set, or several models of monitor (with color as an option). These therefore often had the form factor of a keyboard, such as the ZX Spectrum, or a wedge-shaped box, with the keyboard in front and a flat box behind it that could hold expansion cards and support the weight of a monitor on top, such as the Apple II. The details were often determined by FCC regulations on the allowed level of electromagnetic interference. These machines would therefore need to be redesigned for each country they were sold in (although it wasn’t uncommon for poorer countries to end up with surplus English-language computers).

Apple has made the most variations on this in its Mac product line, with most of its models building the CPU and the monitor into a single unit, and allowing the user to plug in any keyboard (or, these days, use a wireless one). The standard, however, became to make both the monitor and the keyboard separate components that connect through a standard connector, and can be left out if a computer only needs to communicate through a serial or network interface.

Early game consoles used cartridges and did not come with keyboards. The most convenient way to load a cartridge was on the top, but Nintendo of America deliberately opted not to do this with its NES, instead choosing a front-loader design with a spring. Masayuki Uemura claims that this is because they worried that the dry climate in parts of the U.S. would cause the cartridges to build up static electricity.

Another set of interesting physical designs belonged to supercomputers. Seymour Cray designed his multi-million-dollar machines to be cryogenically cooled to run as fast as possible, with some models even conducting heat to an aluminum bar underneath that was immersed in coolant. The Connection Machine was another with a unique design (which looked great on screen), partly because of its large number of distributed processors, partly for the aesthetics.

Most modern computers are hand-held, but only pocket calculators were sold at that size in the time period you’re asking about. Graphing calculators eventually became 8-bit computers with a monochrome display, buttons in front, Z80 CPU and BASIC in ROM, much like the Apple II+ or original Game Boy, but scientific or business pocket calculators had to wait for low-power CMOS electronics to become cheap enough to make this economical. These are also the last survival of that type of BASIC microcomputer, due to regulations on what type of calculators are allowed on standardized tests.

Answer 6

Trust.

In early days of computing, engineers needed to trust that what was entered into the computer was 'in' the computer.

Early Transistors could be seen, you could physically 'see' if they were on/off/blown. This permeated the whole design philosophy for years, those years that you're researching. The Transistors got replaced, but they put little lights on panels to display the needs. Other answers here go into further detail about multiple failure points, each with its own little display light.

As trust was gained by users, we didn't need as many lights, and as the screen was implemented that could 'access' all those data points, we no longer needed panels of lights.

But, It is a single point of failure. If a screen goes down, how do we see all the bugs to be able to fix them. On a spaceship, that has no access to a local computer store, how can we possibly replace a broken screen? We'll want to return to some level of debug via lights, or communicate via ASCII (The Martian) at some point before we can go forward with independent computing.