One thing I hate about US keyboards is that they don’t let me type my French special characters by any memorable strecth – what’s the dead key for the cedilla again? (ç)
One thing I hate about French keytboards is that they have all the punctuation keys messed up… which is particularily painful when trying to write code. < and > share a single key! [ and ] require THREE fingers! (you’d better give un on using arrays!! :p)
But you can switch to use a different layout, such as a French keyboard layout even if you are using a U.S. You can add multiple keyboard layouts and easily switch between them.
The only bearable solution in the long run is to remap the keys to an order that makes a little more sense.
I just found the perfect tool for create new keyboard layouts on the mac: Ukelele ! :)
Ukele easily lets you reassign characters to keys with any hot key combination and it also lets you create dead keys (multiple keypress sequences to create one character).
Now one question remains: which keyboard should I start with: a French AZERTY or a US QWERTY keyboard? (All mac keyboard layouts here)
By the way: if you’re looking for a keymap editor for Windows, here’s one from Microsoft.
By OJ Bucao, September 6, 2010
Being a programmer, I type a lot and I suffer from Repetitive Strain Injury(RSI) and tendonitis on my wrist. I’ve tried many different ways to help makeit better. One way to do this is to switch to a different keyboard layoutother than QWERTY. QWERTY wassupposedly designed for typewriters to solve a very specific problem–tokeep the types from jamming against each other. The most frequently used keyswere placed apart from each other to prevent them from jamming. This resultsin a non-ergonomic layout. However, there are alternatives.
The first alternative keyboard layout that came to mind isDvorak. It wascreated in the 1930’s and promised to be vastly superior to QWERTY.I went ahead and tried it out and soon enough after doing “ls -latr” on theterminal, I had to shake my head and sadly walk away from it. I didn’t likethe way Dvorak was laid out especially for the weak fingers of the righthand.
Then I stumbled upon a layout called Colemak, arelatively new player in the game compared to QWERTY and Dvorak.It was released in 2006 and boasted impressive metrics in terms of fingertravel, hand alternation, and same finger frequency. Everyone in thealternative keyboard layout crowd seemed to be raving about it.There are other layouts available namely Capewell, Arensito, Carpal X, etc.After some research I decided on Colemak because of its metrics and probablypartly because it looked “normal” and “familiar”. The other ones eitherlooked too radical and different or they suffered from awkward placementsof some often used letters. Colemak looked the most promising and I wasexcited to try it.
So I went ahead and tried it and immediately it felt good. I noticed thatmy fingers were not moving up and down as much and most of the timethey stayed on the home row. However after a few days of practicing onK-touch, a nagging feeling started to creep in. Something felt ratherawkward. At first I thought that maybe I just wasn’t used to it yet andit’s the result of the awkwardness in switching to a different layout.So I kept on and while doing so, I tried to analyze how my hands weremoving and then the problem became clear to me.
The Colemak keyboard layout
My initial excitement turned to disappointment when I realized thateven though my fingers were not moving up and down as much, they weremoving too much laterally. I realized that the main culprit wasthe letter ‘H’ placed to the right of the letters ‘N’ and ‘E’.‘N’ is where your index finger rests. Typing ‘HE’ forced the handto make a very unnatural sideways twisting motion from the wristand then back again. To give you an idea on why this could beserious, consider these:
Just ask yourself, how often do you type ‘the’, ‘these’, ‘them’, ‘when’,and ‘where’, etc. on a day-to-day basis? It’s even worse when you’retyping these words in the beginning of a sentence. Try typing “The”with the T capitalized on Colemak and hopefully you’ll see what I mean.Your right hand will move somewhat like this: you swing to theright to get the SHIFT key with your pinky, then you swing back to theleft to get the letter ‘H’, and then you move to the right again toget the letter ‘E’. All this is happening in a split second. That’squite a bit of lateral movement. Now I’m not a doctor, but multipliedover a lifetime, making that sideways motion with the wrists could bedetrimental to people’s hands. It’s nothing personal against Colemak.However, I consider this to be a major flaw in Colemak’s design andI’m concerned that nobody is talking about it. And even if it’s provento be benign, I find it personally cumbersome.
The letters ‘D’ and ‘H’ are relatively high frequency letters and placingthem in the middle of the keyboard forced the hands to make thatlateral twitching move a lot. This is by design since the purpose wasto optimize the home row keys for high frequency letters to reduce fingertravel, which is primarily caused by moving up and down above andbelow the home row. Colemak by design, as well as Dvorak tries toreduce use of the top and the bottom rows. Actually, when you thinkabout it, most of the other alternative layouts optimize for this verything. However, I believe that the way that alternative layouts focus onjust the home row for optimization is somewhat misguided. We shouldoptimize the keys inside the hand’s natural range of motion and notjust strictly the home row.
Other letters that I think are cumbersome with Colemak are the lettersG, L, and O. I believe that by moving these letters, horizontal anddiagonal stretching could be made less and the load on the rightpinky could be reduced.
I was really disappointed that Colemak was not the layout that I hadhoped it would be. I no longer wanted to use QWERTY. I didn’t likeDvorak, and the other alternatives didn’t look very promising either – butrather very alien. I really wanted Colemak to work however I can’tlive with the H-E movement and having to reach for D and H often. Ifelt that it could be made better.
I tried to see if there’s anything that could be done to solve this.At first I ignorantly tried to replace ‘D’ and ‘H’ with other lowerfrequency letters and moved them elsewhere still expecting the samemetrics. I used the awesome Keyboard Compare applet by John A. Maxwellwith modifications from Michael Capewell, and also Patrick Gillespie’samazing Keyboard Layout Analyzer. Long story short, I got prettycrappy results. It soon dawned on me that just moving a few thingsaround isn’t going to cut it. It’s like playing with a water balloon.If you squeeze on one side, it bulges on other sides. If I was goingto get the results that I’m looking for, I had to sit down and do somethinking.
I decided to try to create a new keyboard layout based on theseideas. I first came up with the following observations and assumptions:
Most of these seem obvious enough but it helps to jot them downfor clarity. I then came up with a set of principles to serve asguidelines to help me with the design:
Here’s a an illustration that I created grading the keys based on theapproximate amount of difficulty/strain in reaching or pressing themwith 1 being the easiest and 5 being the most strenuous. This gradingscale takes into consideration the position of the keys, the strength ofand length of the fingers and the staggered nature of the keyboard.
Keys graded based on strain/difficulty (Standard Keyboard)
Below is what it would be on an “matrix style” keyboard alsoknown as “grid” keyboards.
Keys graded based on strain/difficulty (matrix style Keyboard)
I call it the Workman Keyboard Layout in honor of all who type onkeyboards for a living. And considering that today is Labor Day, I thinkit’s perfectly fitting.
The Workman Keyboard Layout
In Workman-P, the top-row numbers and symbols havebeen switched as well as the brace and brackets. It is great forprogrammers as well as system administrators.
Workman for Programmers
On Workman, the most often used keys are evenly and pleasantly distributedinside the natural range of motion of the fingers. It’s even betteron a matrix style keyboard.
Workman reduces overall usage of the two middle columns by about50% over Colemak. This 50% reduction can be divided into two parts,horizontal and diagonal index finger stretching. Workman reduceshorizontal finger stretching by 63%, and diagonal index fingerstretching by 27% over Colemak. This is because Workmanefficiently utilizes other easy to reach keys instead of justplacing them in the middle columns where they are difficultto reach. Workman also reduces vertical index finger stretchingby 30% over Colemak by realizing that it’s easier for the indexfinger to fold than reach upwards.
Below are some tests using popular books taken from Project Gutenberg:
Looking at the first example. Colemak achieves the lowest overallfinger-travel distance against QWERTY and Dvorak at 30,352 meters.However, Workman is even lower at 29,656 meters — a differenceof 696 meters. It doesn’t sound like much, however if we convertit to centimeters, that’s equal to 69,600 cm. And consideringthat the distance between keyboard keys is approximately 2 cm,typing on Workman is like typing 34,000 less keystrokes thantyping on Colemak. At 40 words per minute, that’s equivalent toapproximately 3 hours of work.For Dvorak, it’s 126,000 keystrokesat 11 hours of work. And for QWERTY, it’s 1,369,800 keystrokes at 5days of work.
This shows how many times you had to do a double combo with oneof your fingers. For example, typing the word “fuel” using Workmanmakes your right middle finger do a double combo because the lettersU and E are both typed using the right middle finger. Here, Workmanhas an SFU of 2.185% which means that for every 46 keystrokes(approx. 9 words), one of your 8 fingers does one double combo.Compare that to QWERTY which is at every 20 keystrokes (4 words).Colemak is at every 58 keystrokes or (11 words). Workman, onaverage, has a higher SFU than Colemak… at +1%. Some peoplemisunderstand and think that this somehow shows increased effortor discomfort. It doesn’t. Effort is the same, because no matterwhat, you’re still pressing the same number of keys. Comfortshouldn’t be a problem as long as the key is in a comfortable spot.The only thing that SFU might potentially and theoreticallyaffect is speed because typing two letters with differentfingers is a little faster than typing them with the same finger.However, I doubt that most people will have any problems with speedat all using Workman especially considering that very many peopletype very fast on QWERTY, of all layouts.
In case you were wondering, the bulk of Workman’s SFU comes fromthese combinations: LY, OP, PO, CT, and UE. All ofthese combos are very comfortable to type with LY being lesscomfortable because the movement from L to Y is diagonal. Somepeople might say that this is a very bad thing but in reality it isnot. First, LY occurs at about 0.24% of the time on average.That’s less than a quarter of one percent. To put it into perspective,for every 10,000 keystrokes, you will type LY only 24 times.At this rate, you will not even notice it. Even with this extra0.24% considered, Colemak’s diagonal movements are still greater thanWorkman’s. Second, even though it’s a diagonal motion, you’re notreally stretching that much because when you type L, you foldyour fingers (storing potential energy), then you release it totype Y. The stretch is about the same as when you come fromhome row. It’s even less when you use a matrix style keyboard.Third, LY occurs at the end of the word almost all thetime. This is important and it makes a huge difference.This means that when you type LY, you do it at the end ofthe flow of a word as a finishing stroke instead of beingin the middle, which makes it less cumbersome. All in all, Idon’t think this is a big deal.
A better indicator of finger effort is the Finger Percentage. Ifyou look at the Finger Percentages for Workman, Colemak, and Dvorak,nothing really stands out at first glance. However, Workmanfurther reduces the load on the right pinky finger over Colemakand Dvorak. The right pinky, despite being one of the weakest, isone of the most used finger on a standard keyboard due to thelocation of the Enter, Shift, and Backspace keys, as well asadditional punctuation keys. Both Colemak and Dvorak have higher rightpinky percentage at 11% (253,850 keystrokes), while Workman isonly at 9% (207,696 keystrokes). On Workman, your right pinkyfinger just typed 46,155 less keystrokes than both Colemak andDvorak… that’s about 4 hours of work using ALL your fingers.
Below are the average percentages for each hand. The two analyzersgive slightly different results because they differ a little bitin how they do the calculations. However you still get the idea.QWERTY has about a 4% lean towards the left while Colemak leans tothe right by about 5%, and Dvorak, 7%. Workman balances theload between the left and right hands almost equally at 50%.
Patrick’s Analyzer | John’s Analyzer | |||
---|---|---|---|---|
Layout | Left Hand % | Right Hand % | Left Hand % | Right Hand % |
QWERTY | 54 | 46 | 53 | 42 |
Dvorak | 44 | 57 | 43 | 54 |
Colemak | 45 | 56 | 43 | 54 |
Workman | 50 | 50 | 49 | 48 |
Dvorak consistently gets lower Same Hand Utilization than QWERTY,Colemak, and Workman which are usually in the 30% range whileDvorak is in the 20’s. Dvorak was supposedly designed for low SHU whichmeans that your hands alternate more frequently. Dvorak’s 20% SHUmeans that on average, you’re typing 8 keystrokes alternating betweenyour hands, and the next 2 keystrokes, all in one hand as a combo.30% SHU then means on average, 7 keystrokes alternating andthen the next 3, all in one hand as a combo. In designing Workman,I preferred a high SHU (low alternation) over a low SHU (high alternation).I think high alternation is beneficial if you’re typing onmechanical typewriters but not necessarily on modern keyboards.On typewriters, it is very difficult to type combos with one handbecause each key needs a large amount of force to depress. Youactually rely more on the momentum of your arms and wrists toprovide that force so alternating between your two arms isvery helpful. However, this method of typing is inefficient onthe modern keyboard because modern keys are easy to press.You are no longer reliant on each arm or wrist stroke todepress a single key. Doing so is actually unnecessary and awaste of energy. It is much more efficient to ride themomentum of a single arm or wrist stroke and type a combo ratherthan just one key. This way your arms and wrists potentially moveless while typing the same number of keys, effectively killingseveral birds with one stone. In the beginning, this will not beapparent. However, as you become more proficient and familiar with thecombos, you will be better able to utilize this advantage and typebursts of familiar texts in one hand using fewer hand strokes. Anexample of this is the word OPERATION. If you were to typethis in Dvorak, you could type it as o-pe-r-a-t-io-n whereeach grouping is a hand stroke–a total of 7 hand strokes.Whereas with Workman, you’d probably be able to type it aso-pe-rat-ion using only 4 hand strokes. Typing Don Quixote,your wrists and arms potentially moved approximately 200,000 timesless on Workman than on Dvorak.
What these stats do not show is the usage of the middle two columns.Colemak puts 280,850 keystrokes (12%) on the middle columnsversus Workman at 125,875 keystrokes (5%). On Workman,your index fingers (and potentially your wrists) movedsideways 154,975 times less than on Colemak.Dvorak is at308,533 (13%) and QWERTY is at 512,568 (22%).
I encourage you to do your own testing and analysis. Notethat different keyboard testers will give different results asto what layout is better depending on the criteria that theyare using to do their measurements and assessments. SinceWorkman’s philosophy is unique, many testers will registerit inferior to others.
To do your own testing, you can use Patrick Gillespie’sKeyboard Layout Analyzer.
You can grab full texts of public domain books here atProject Gutenberg.
Sure go ahead! Feel free to use it if you would like. Below is alink to the implementation/installation files courtesy ofDavid Norman (deekayen).
IMPORTANT: The Workman Keyboard Layout is only a partialsolution. Even the best keyboard layout could not completelyremove the risk of typing injury. Typing in itself is anunnatural and hazardous task and no keyboard layout couldprevent injury without proper precautions and common sense.I suggest learning to type with good hand and finger posture,taking frequent breaks, keeping your hands and wristswarm while typing, and using a keyboard that meets your needs.Our health, after all, is ultimately our personal responsibility.
I hope that you’ll enjoy this layout and benefit from it. Ifyou like the Workman Layout, feel free to tell others about it.