Informatik Handwerk
Peter Fargaš
Programmer :: Prototyping, Research
PHP | JavaScript | Java
Informatik Handwerk
Peter Fargaš
Programmer :: Prototyping, Research
PHP | JavaScript | Java
Informatik Handwerk
Peter Fargaš | Programmer :: Prototyping, Research | PHP,JavaScript,Java
Release date: September 2015
Link to authoritative version https://knowledge-transfer.informatik-handwerk.de
/article/inspirative/dvorak/unreleasedDraft.html

A Modern Keyboard Layout:
wide, prg, num, dead, dvrk, rl

On this document

The most important part, something almost by a rule missing in other resources available, is the last section which describes my real-life experiences and observations.

I am not exhaustive within this document - I resisted to reformulate and include info from linked resources. I serve the important bits as (partial) mirror images. Apart from that section being a collection of links, it is as well something to read, not scroll through.

Motivation, Conclusion

Most probably, keyboard will be my main tool for the rest of my life, so I decided to invest into research. Layout alteration, is a piece of that puzzle.

If you are driven more by the important and less by the practical - a fresh layout is a wonderful exercise and lots of fun!

I myself desired a different layout for symbols - they are hidden beneath numbers or so far on the side, that typing them made me either loose the track of where my hands were positioned or I had to look at the keyboard. As I was already at it, I tried out, tweaked, dvorak.

Unless your usage of keyboard is highly specialized, like for example programmers have it, I don't see a point in switching. But if you wish to get into deeper materia, I consider it an unavoidable step. I find the variations available are largely equivalent and their offer rather weak. As a part of a larger landscape, there are key issues seldom taken into account (appearing only inderectly in this article).

I find the time is not ripe for switching to new strain of keyboard hardware - grid keyboards, chorded keyboards, etc. Available alternatives are awesome (like the open steno efforts), yet there is plenty of ground to discover and collectively test on the standard, simplifying promotion of efforts to evolve this kind of, single-glyph, input devices.

Established standards from typewriter times

QUERTY
  • 1873 design
  • early typewriter was in need of a speed limit (the design jammed while typing too fast)
  • 32% letters typed on home row
  • I disagree with, there is no evidence: the inefficiency projects onto chronic problems with shoulders and wrists
  • slight bias on left hand
Prof. August Dvorak
  • 1932 design
  • a decade of work at that time
  • each finger is utilized proportionally to its skill
  • 70% typed on home row, 22% upper, 8% lower
  • successive keystrokes fall on alternative hands
  • slight bias on right hand
  • the letter grouping is much more intuitive to learn

What is to be kept in mind are the very different mechanical properties compared to modern keyboards: literally "hitting" instead of "pressing" keys.

External resources

deskthority.net - Keyboard layouts
wikipedia.org - Keyboard layout
Listing most common variations - obligatory mention, I guess.

Minimak Layout
Colemak Layout
Arensito Layout
Aset Layout
Norman Layout
Design argumentations, comparisms. Depending on the cost function, finger travel distance shrinks up to 50% on english prosa. The relevance of those metrics are disputable. The typing speed increase or lerncurve get no attention. Dvorak is estimated to offer 15% gain after months of usage and I don't think others get more. I classify all as rather equivalent contributions, even though here and there some exotic can be found, like capslock-backspace swap.
Workman Layout
Capewell Layout
Workman layout is pretty serious with the story how the layout evolved, Michael Capewell stands out for the broadness of his explorations.
A Programmer's Keyboard Layout from Kyle Lahnakoski
I most like the contribution of Kyle (clean minimal structured talk, no promo) starting with an inspirative list over design modalities, great sollution approach and a good, sharp conclusion I agree with. Don't miss, here are the local pdfs of list over design modalities and conclusion.

carpalx project
Algorithmical search and evaluation of layouts. Neglects the relevance or superiority of alternatives to QUERTY - they are primary remedies for people with repetive strain injury.

Daniel Andersson's opinion on layouts
Explaining layout variations in broader context, exposing the cons and directing attention towards more relevant concearns (well known by the industry). I am including this as a local pdf.

Experiences, Observations, Recomendations and Guidance

Typing Styles

  • On QUERTY, my style of typing was "free-floating hands with loosely assigned fingers" and for the new layout, I wanted the classical "FJ-locked 10 finger system" approach. I believe I can consider myself lucky - larger distances and quirkier movements on QUERTY v.s. the smoothness of Dvorak do suggest different choices of typing systems. Most probably, the brain can couple processual and spatial systems with less confusion and even keep both actively, colisionless available.
  • Thus, I consider natural learning of QUERTY as optimal and not being a touch typist for that layout as superior. The "hunt-and-peck" style, once mature, can be very easily upgraded by starting to trust the invested time, and gradual incorporation of other fingers, eventually reaching speeds on par with touch typing.

Learning

  • I used diy-stickers which were peeling off and they (and that) did great.
  • Switch learning techniques regulary. Takes some time. Don't give up - it is a wonderful excercise.
  • Common typing style for different layouts is potentially problematic.
  • Small change once in a while are easy to get used to.

Integration with existing software

  • Software can use 3 different kinds of bindings - the hardware "ScanCode", the (technologically somewhat premature) conceptual "VirtualKey" and the one you already (could) know, the language-layout. The hardware is unmodifiable, the language-layout is the one you are redefining. You can get dirty with VirtualKeys only to a certain extend and I believe it complicates matters.
  • Good news is that the shortcut keys are mostly aligned on ScanCodes or VirtualKeys, so you often keep "the space of your combos". On some applications, you will have to get used to new locations however. (N.B. I consider the choices on shortcut "letters" in almost every application as faulty design decision)

Layout Design Basics

  • Hand optimized layouts are comparable to algorithmically generated.
  • The great thing is, that wherever you come, the keyboard has the "abandoned" standard on it - so even after a longer time, you can not get lost out there.
  • Be cautious in decision on the new layout, at least in the coarse. I swapped hands after "getting almost there" and it was a big challenge on my will not to throw it away alltogether. Even now, I still experience residua now and than.
  • For the insight gained from integration, maximum compatibility, I suggest keeping ZXCV(undo-cut-copy-paste) at old locations.
  • If, than bias your layout on the left hand - the right hand already carries Enter, Backspace, Arrows, Mouse.
  • Experiment and hunt down awkwardities which bug you. Small change once in a while are easy to get used to.
  • My opinion on alternating hands is, despite general opinion and the correct argumentation of Kyle Lahnakoski, towards using it. Training of hemisphere cooperation is much more important than some theoretical, marginal speed improvement. I recommend tweaking Dvorak.

Installing new layouts

  • For Windows, "Microsoft Keyboard Layout Creator" does great.
  • Did no research for Linux or Mac.

My variation: ergonomy and programming

  • At the time I started, I was not aware there is such a plenthora of available layouts - my base was Dvorak. I swapped hands, kept ZXCV, rearranged symbols.
  • I have widened the gap between hands, the base for the right hand is the former K-key (I here). More ergonomical, better access to enter, backspace & arrows plus the benefit of central area.
  • In programming, symbols are more often used than numbers - I swapped the shift on this row and split the numbers to maximize optimal access. The arrangement of symbols is my personal taste - groupings of logical, negation, arithmetical, spatial symbols. Plus some other at convinient locations.
  • My layout expects a 102-key keyboard, became the Q-key. Was supposed to be where I considered underscore more useful.
  • Since the 102-key keyboard has space for a dead key, there it is - in the (awkward) middle.
(mouseover for shift-states)
layout as presented by Microsoft Layout Creator