The keyboard is the eldest peripheral device—it is in fact about a century older than the computer itself. The average office worker spends 8 hours a day pressing the keys, still we pay surprisingly little attention to the quality of our keyboards.
If we want to turn to ergonomic keyboards, our task is neither simple, nor cheap. As for me, I built one for myself.
My story began about one and a half years ago, when I developed tendinitis first in my right wrist, then in the left as well, which hindered my typing as well as every other activity I do. I started looking for an ergonomic keyboard at once, however, on the Hungarian market I only found 1-1 product from two brands. I bought one of them and my tendinitis slowly disappeared. Still I could not let the issue lie, so I continued reading keyboard-related forums and stumbled upon several interesting things.
The average keyboard
The ancestor of keyboards, the typewriter was first built around 1870. The still subsisting QWERTY layout was developed by Christopher Latham Sholes in 1973. He spaced out he most commonly used letters (in English) in order to prevent type bar jams. In other words, the original goal of this layout was to slow down typing. (Fun fact: the top row letters include all necessary characters for typing the word TYPEWRITER, which was probably arranged so on purpose.)
We are still using this artificially decelerated layout, although typewriters with type bars are quite rare today: instead of levers, today's average keyboard has 104/105 rubber domes and a switch matrix on slim plastic membrane. When pressed, the domes collapse and close the electric circuit by compressing the membranes at given points.
Some love the feel of how a rubber dome collapses and the key gives a crisp „plastic” snap against the membranes. In case a problem arises - e.g. tea or coffee drips among the membranes, the replacement is quite cheap.
Many of us are averse to this type of keyboard.
On top of it all, you cannot properly solder diodes to the membrane, so if the circuit is closed at a given point, then electricity also flows backwards through the entire row or column; thus in case keys are pressed simultaneously in an unfortunate combination, ghosting might occur, meaning that a fourth key press is incorrectly registered. Firmware often handle this problem by simply ignoring suspicious activity—this is why gamers sometimes encounter the phenomenon that the computer only senses a few of the 3 or 4 keys pressed simultaneously (e.g. jumping from a diagonal crouch).
… there is better!
The simplest and most accessible solution is to use an alternative layout.
The best known and most widespread alternative is the Dvorak, which was developed by Dr. August Dvorak in 1936, based on the frequency of use of English letters. On average, 70% of all strikes are in the home row (compare to QWERTY's 36%), vowels are all on the left (two hands type faster in turns), there is a version with Hungarian characters as well, and it is available almost everywhere.
But, as only A and M coincide in QWERTY and Dvorak—the frequently used ctrl+c/ctrl+v is also found elsewhere in the layouts—several other alternatives have popped up recently. These also have nice statistics, like Dvorak, but take more points of view into consideration—e.g. they leave more letters of QWERTY in their place. Such are the Colemak and the Workman layouts, beside innumerable others.
These alternative layouts are without doubt great, still there is one major problem with them: if you can do 10-fingers typing you'll have to re-learn it—if you cannot, it is improbable that your wrist hurts on account of typing—and once you memorized the new layout, you'll meet QWERTY everywhere else and find it difficult to use.
Changing the layout does not solve the primary problem that typing on a traditional keyboard means our wrists are unnaturally twisted, while the keys appertaining to given fingers are also staggered.
This is why ergonomic keyboards were developed, some of which are available in Hungary at a not-so-cheap price, but are only moderately ergonomic (see the above-mentioned models). Others are more ergonomic, but heavy, strange-looking, and surprisingly expensive—also, they have high delivery prices, as they are practically unavailable in Europe.
Here are some of the most popular models: Truly Ergonomic, Kinesis Advantage and Kinesis Freestyle. Unfortunately, I have not had the opportunity to test either one.
At keyboard forums, people do not only chat about keyboards, but also design and construct new ones—because some of us want a one-of-a-kind keyboard, while others are simply dissatisfied with the commercially available keyboards. On top of it all, the majority of these projects are open source, so you can also help smooth out any edges or build your own custom model.
Before building a keyboard, let's see the parts of a classic, high-quality mechanic keyboard, which is unlike today's mass products.
The printed circuit board
This is where you find the matrix connecting the switches—with the addition of diodes to prevent ghosting—, the controller, and the LEDs. The switches are attached here by soldering as well.
Models that are smaller than the traditional keyboard, like the TKL (tenkeyless; has no numpad) and the 60% keyboard are quite popular—they take up less space, and also need less switches and keys, which are not cheap.
The switch plate
In order to relieve the printed circuit board of all the pressure that comes with pressing a key, the switches are placed onto a rigid metal or plastic plate. This tray determines the final position of keys as well—the printed circuits usually allow for various layouts (ANSI - American, slim enter; ISO - European, L-shaped enter, additional key in place of “í”; etc.)
We could also discuss mechanic switches at great lengths, since they can work in various ways and they are by no means technological novelties. The currently most popular Cherry MX series, for one, has been on the market in unchanged form since 1985.
If your printed circuit board only supports a given contact mechanism, you still have numerous options to consider - e.g. by actuation, key feel, etc.
Choosing your keycaps is not a simple task either, since you have to decide about the profile (SA, DCS, DSA, etc.), colour, material (different types of plastic and metal), legends, and keycap printing method (printing, laser marking, double shot injection moulding, etc.).
You have several options for acquiring a keyboard case: re-use and old one, mill one from wood or aluminium, assemble from laser-cut acrylic sheets, or 3D print one. Plans and designs are available on the internet.
In the majority of the open source DIY keyboard projects that I've seen, controlling was usually entrusted to Arduino-compatible devices, thus the standard gnu-avr toolchain was at disposal for firmware development. Of course, firmware are also open source and customizable.
So I decided to build myself an ergonomic split keyboard, the ErgoDox.
I procured the necessary parts through keyboard forum group buys, which meant I could save significant amounts of money. (The commercial price of a simple Cherry MX switch might be as high as 1 EUR, but in case of an order of thousands of pieces the price could well be halved.)
I spent about 1 hour per day on assembly, thus it took me about 6 days to finish. I have added mouse and NKRO support to the firmware, which means that my keyboard is capable of generating mouse movements and clicks, while it is also capable of transmitting theoretically infinite simultaneous key presses - due to the incomplete USB HID implementation of some BIOS-es it is only possible to transmit 6 key presses without using tricky transfer methods.
Besides having acquired an ergonomic split keyboard at a lower price than the aforementioned expensive models, I could also refresh my soldering and firmware programming skills, so the project was well worth the effort. I am typing this article on the very same keyboard, on a DVORAK layout, at a relatively slow speed for the moment. :)