Alternative keyboard layouts

From Deskthority wiki
Revision as of 07:07, 23 October 2017 by Yellowfour (talk | contribs) (BEAKL (2016): link to main article; update to BEAKL 9)
Jump to navigation Jump to search
Template icon--Rethink.png This article requires restructuring

Modifications of QWERTY

QWERF (2006)

Highlighting changes from standard QWERTY.

QWERF JYKL;
ASDTG HUIOP
ZXCVB NM,./

http://www.michaelcapewell.com/projects/keyboard/index.htm#The_QWERF_Layout

QWERF is a layout designed by Michael Capewell for ease of learning and great efficiency-boost. It supersedes his earlier Capewell-QWERTY (C-QWERTY) and an earlier iteration of QWERF.

Claims 28% finger-movement reduction, and 38% if E/D are swapped too, though in that case Capewell warns against potential discomfort while typing of bigrams ER and RE.

CarPalx Partial Optimisation

The 5-swap (QWKRFY) layout

QWKRF YUIJP
DSATN HOEL;
ZXCVB GM,./

http://mkweb.bcgsc.ca/carpalx/?partial_optimization

The carPalx project, authored by Martin Krzywinski, created a "Partial Optimization" to see the maximum efficiency-boost for the least amount of key-swaps.

A listing of the best individual keyswaps is included. The top 5 swaps are as following, with effort reduction as calculated by carPalx in parenthesis:

  1. K/E (-13.4%)
  2. J/O (-11.1%)
  3. F/T (-9.0%)
  4. D/A (-5.8%)
  5. G/N (-5.6%)

Aggregate effort-reduction is 37.7%. The semicolon was ignored, the first key to change position in other layouts.

CarpalxQ (2007)

CarpalxQ layout, highlighting changes from standard QWERTY.

QWKRF YULP;
ASDTG HNEIO
ZXCVB JM,./

http://jumpedthesynapse.blogspot.com/2007/07/carpalxq.html

CarpalxQ is a layout created by Jay Walker, as a modification of a no-longer-published iteration (called Mod-1) of carPalx partial optimisation. It consists of the following six key-swaps:

  1. E/K
  2. T/F
  3. N/J
  4. I/L
  5. O/;
  6. ;/P

Martin Krzywinski talks about it and compares it with his partial optimisation: http://mkweb.bcgsc.ca/carpalx/?partial_optimization#carpalxq_user_contribution

The home-row top-row partial swap layout

Qwerty-rowswap.png

QWDFG YJKL;
ASERT HUIOP
ZXCVB NM,./

Credits to Jammycakes from the Colemak forum? http://forum.colemak.com/viewtopic.php?pid=7152#p7152

To understand the changes from QWERTY, try to trace the QWERTYUIOP string!

This layout swaps a few keys strictly vertically so as to keep everything under the same fingers. This should achieve a very simple transition from standard QWERTY.

A carPalx analysis has been done on the layout:

Total effort: 2.122 (vs. qwerty: 3.000, dvorak: 2.098, colemak: 1.842).
Effort contributions: base 0.491, penalties 0.743 (0.217 row, 0.408 finger), path: 0.889.

7BIT Layout (2011)

7BIT-Layout is a layout which is very similar to ANSI/US-QWERTY, but with strong influences from the Lisp/Symbolics and Emacs culture. True 7BIT layouts have got a short space bar (1.5 units) and as many as possible modifier keys in the bottom row. colon and backquote are swapped, Caps Lock, if present at all, is moved to the function key row and at its usual place (where Control on Unix keyboards is) is a modifier key.

Layouts based on a standard keyboard

Dvorak (1936)

',.PY FGCRL
AOEUI DHTNS
;QJKX BMWVZ

Soon after QWERTY was born, its incredibly poor design lost all justification, but typewriters continued to be built with it. Dr. August Dvorak set out to create a superior alternative. He gathered data in the form of statistics and closeup films of typists using their fingers. That research created the Dvorak layout.

  • The most common letters were roughly placed in the rows typists were thought to like best: home-row, top-row and last bottom-row.
  • Hand-alternation: To avoid occasional long strings of one-handed-operation, emphasis was placed on hand-alternation. All vowels were moved to the left side of the keyboard, and onto the home-row, including the rare U.
  • The right-hand is favoured and receives 14% more usage than the lefty, which is 33% more than QWERTY because it favours the left hand with a similar percentage.

Criticisms of the Dvorak design:

  • Extreme dissimilarity from QWERTY. This may be a good and bad thing – with dissimilarity some argue that it will be easier to tell the two layouts apart. However, learned hotkeys (e.g. Ctrl+C) may have to be relearned, and user-experience with certain applications may be ill-affected.
  • Hand-alternation. This is argued by some to be counterproductive. The computer keyboards of today are far lighter to operate than typewriters, and it may be argued that preference has therefore shifted to pressing several keys in a row with the same hand.
  • Poor placement of common letters. L is a case-example of this, receiving many complaints among disillusioned Dvorak-users, in particular those running Linux.

Programmer Dvorak

http://www.kaufmann.no/roland/dvorak/

Developer's Dvorak

See also Advanced Developer's Dvorak.
* Standard: https://web.archive.org/web/20130512063159/http://www.andong.co.uk/dvorak/DDvorak.aspx?page=Standard+DDvorak
* Light: https://web.archive.org/web/20130512065149/http://www.andong.co.uk/dvorak/DDvorak.aspx?page=Light+DDvorak

Neo (2004)

A picture of the multiple layers of the Neo-Layout

xvlcw khgfqß
uiaeo snrtdy
üöäpz bm,.j

http://neo-layout.org/ (page is in German)

The Neo-layout was developed with the experiences of other ergonomic layouts like Dvorak in mind. It is mainly targeted at German users, but supports nearly all characters of Latin-based alphabets, as well as the Vietnamese and some African alphabets. It consists of six layers, available by combinations of multiple modifier keys. Using these layers there are special characters available, which are not, or not as easily available in QWERTY-based layouts, like mathematical symbols, and greek letters. Every Layer serves a special purpose:

Layer Usage
1 Lowercase characters
2 Uppercase characters, typographical characters
3 Special characters, especially for programming, etc.
4 Navigation keys, Numberblock, etc.
5 Greek lowercase characters
6 Mathematical symbols and Greek uppercase characters

Capewell-Dvorak (2004)

',.PY QFGRK
OAEIU DHTNS
ZXCVJ LMWB;

http://www.michaelcapewell.com/projects/keyboard/layout_capewell-dvorak.htm

A modified version of Dvorak designed by Michael Capewell.

Attempts to solve many perceived faults with the Dvorak layout, such as:

  • Placement of the L letter
  • Swaps I/U because I is much more common than U and should be nearer
  • ZXCV return to their original QWERTY positions

Capewell (2005)

.YWDF JPLUQ
AERSG BTNIO
XZCV; KWH,'

http://www.michaelcapewell.com/projects/keyboard/layout_capewell.htm

Not recommended by the author; a dead work-in-progress. Nevertheless interesting to read about.

Colemak (2006)

This image compares QWERTY and Colemak.

QWFPG JLUY;
ARSTD HNEIO
ZXCVB KM,./

http://colemak.com/

Colemak is an alternative keyboard created by Shai Coleman, named as a portmanteau of Dvorak and Coleman. Its design goals consist of easy transition from QWERTY due to repositioning only 17 letter keys. Additionally the AZXCV shortcuts are in the same location perhaps allowing an easier time switching from QWERTY.

It also claims greater efficiency than Dvorak. Furthermore it places complete emphasis on the home-row: the ten most-common characters in English are on the ten home-row keys.

Setting up Colemak

A physical example of a Colemak layout on a Filco Majestouch 105 with Group Buy Round Three doubleshots depicting Caps lock and Backspace reversal as well as a customized United Kingdom layout

The layout is included with most Linux-distros (under the US header) and the Mac OS X 10.7 "Lion" developer versions(outdated info?). Windows-users should go to the official Colemak website to download the keyboard layout installer.

It should also be noted that Shai Coleman proposes to replace the Caps-Lock key with an additional Backspace. The default Windows-install does not remap this key due to limitations with which keys are allowed to move in Windows. However, an additional registry script is included which will remap the Caps-Lock key into a Backspace regardless of the current layout, if you want it.

Custom variation: Backspace and Caps-Lock reversed

The recommended layout disposes of the Caps-Lock key, which is not to everyone's liking.

It is possible to create a Windows setup-file for the Colemak layout for those who want Caps-Lock to take up Backspace's original position. The necessary files available in both US English and UK English can be obtained here: http://www.mediafire.com/?gg3u3o0erpcka

Custom variation: Mod-DH

Some users find the placement of the D and H keys unduly awkward in standard Colemak. To address this, there is a variation of Colemak called Mod-DH.

CarPalx (2008-)

QGMLW BYUV;
DSTNR IAEOH
ZXCFJ KP,./

http://mkweb.bcgsc.ca/carpalx/?full_optimization

The prime layouts developed by Martin Krzywinski using his layout generator under the carPalx project. As of Feb 2012 these are known as the Q*MLW* layouts, which involve three variations:

  • QFMLWY – Only allows the letter keys to move from their original QWERTY positions.
  • QGMLWB – Like above, but also allows the semicolon to move up, yielding a "Colemak letter mask" ("full optimisation")
  • QGMLWY – Like above (QGMLWB), but forces the ZXCV keys to stay in place.

AdNW (2010)

The AdNW-Layout

AdNW ([1]) is a Dvorak-type layout for German and English. Special versions exist for other languages and ergonomic keyboards.

History and name

AdNW is a layout that emerged from online discussions about Neo, a layout that was meant for the German language. The participants agreed that Qwertz/Qwerty was bad for typing German, and that Dvorak was very good (be it stock or with extra German letters öäüß). The heated discussion focused on the question: was this new Neo layout even better than Dvorak? Some participants in the discussion were not convinced it was, and started a systematic search for a single layout that would outperform Dvorak both in English and in German. To do this, Dvorak's criteria were coded in a newly programmed "Optimizer". Since the group formed itself during the Neo-discussions, they called their layout "From the Neo World", which is in German Aus der Neo-Welt, abbreviated to AdNW.

Philosophy and performance

AdNW is based on roughly the same ideas as "Dvorak" keyboard. Being in the Dvorak tradition, the following aspects are important:

  • Same finger use (low)
  • Adjacent finger use(*)(low)
  • Inward motions versus Outward motions (high)
  • Home row use (high)
  • Row jumps (low)
  • Finger balance (less on pinkies, more on middle and index)

(*)According to the adjacent finger criterium, (qwerty) AS and SD are rated as not so good; AD and SF are seen as better. The idea is that adjacent fingers, especially the pinky and ring finger, are not completely independent, making "rolls" with adjacent finger less pleasant and therefore to be avoided. "Rolls" on index and middle finger (like qwerty ER) are less problematic and get a lower penalty for that reason.

As a result of these criteria, the AdNW layouts also have a balanced Left/Right distribution (roughly 50% of effort on each hand, compared to Qwerty that puts most work on the right hand) and a high hand alternation. Alternation means that common letter combinations like ER or IN are not typed on one hand (like in Qwerty) but on two hands. In AdNW, E and I are on the left side, R and N on the right.

To make the differences between Qwertz and AdNW clear, they are compared both visually and statistically. The graphs and data are produced by the AdNW optimizer. In the graphs, the letter 'flow' is shown. The more common a digraph is (say, ER), the fatter the line that is drawn between these two letters.

AdNW and Qwertz compared

Looking at the graphs, one sees that qwerty uses the left hand a lot, especially the top row. Some of the most frequent bigrams are on the left top row (WE, ER, ET, RT). On the right hand the frequent IN bigram includes an home row jump, which is seen as highly unwanted. Compared to Qwerty, AdNW is more balanced (L/R), has more home row use, and much less one handed bigrams. Frequent bigrams (like WE, ER and so on) are not typed with one hand, but include a hand alternation.

Even if AdNW was optimized for 50/50 English/German, it performs very well for pure English as well. These are the statistics.

adnw and qwerty compared

Compared to Qwerty,

  • AdNW is more balanced left/right (52.7 versus 59.0)
  • AdNW is more home row oriented (72% typed on homerow, versus 32.6% in Qwertz)
  • AdNW is more balanced over fingers (less use of right hand index and middle finger; more of other fingers)
  • AdNW has more alternation (70.8% versus 52.2%)
  • AdNW has less use of adjacent fingers(**)
  • AdNW has less same finger use
  • AdNW has much less home jumps

A multi keyboard, multi language layout

From the beginning, AdNW has known several versions, apart from the "standard" version (HIEAO DTRNSß). The other versions place slightly different weight to Dvorak's criteria (such as: even lower adjacent finger use, at the expense of other criteria). Some of these versions are designed for different keyboards (matrix/orthogonal layouts, Ergodox, TEK etc.) or for other languages. A user can also calculate a custom layout.

Reasons for calculating a custom layout include:

  • use of non-standard physical keyboard, for instance a Planck, a Space Cadet, a modded Ergodox or a DIY split keyboard;
  • different preferences regarding finger use, alternation, and so on. Users may sacrifice performance on one aspect (e.g. alternation) in order to gain better performance on others (even lower same finger use).
  • damaged hands, shorter fingers
  • input language. Users may not type 50/50 English-German prose, but for instance 30% French scientific prose and 70% Python code.

All this can be combined: a user may calculate an optimal layout for "30% English prose, 40% Swedish forum use, 30% Polish, for Maltron keyboard, that avoids using the pinkies and that prefers the bottom row over the top row". Because of this freedom, there is not one single AdNW layout; instead AdNW is a multi keyboard, multi language layout.

Here is an example of an AdNW layout for a special "almost orthogonal" split keyboard (The "Red tilt" keyboard, named after the used Cherry MX switches, by Deskthority user "suka").

AdNW Red Tilt

Some other examples of AdNW layouts are:

standard AdNW

KUÜ.Ä VGCLJF
HIEAO DTRNSß
XYÖ,Q BPWMZ

Bu-Tek - for the Truly Ergonomic Keyboard

ßbu.,ü pclmfx
_hieao dtrns
_kyöäq jgwvz

KOY - even less adjacent finger use

K.O,Y VGCLßZ
HAEIU DTRNSF
XQÄÜÖ BPWMJ

AdNW-NL_EN - for Dutch and English - like old Dutch typewriters, it has a dedicated ij key (Unicode 0133)

buy,! fpvmljx
saeio gdtnrw
z:.ij/ kcmhq

Without ß Ä Ü Ö If these keys are not needed, the positions can be used for other symbols, for instance:

KU!.@ VGCLJF
HIEAO DTRNS:
XY/,Q BPWMZ

Comparison to other layouts

In the table, AdNW is compared to several other layouts.

Layout In short User base Integrated in OS ? Optimized for.. Multi-lang Custom versions Main goals(*) Criteria quantified? Optimizer? Online forum ?
AdNW alternating pico X11 German/English yes yes 1,2,4,5,6,7 yes yes yes
Dvorak alternating micro OSX, X11, Windows English no no 1,2,4,5,6,7,8 no no yes
Colemak rolling nano X11 English no no 1,3,4,5,6,9 no no yes, large
MTGAP rolling pico - English yes yes 1,3,4,5,6,7 yes yes yes
Carpalx own thing pico - English yes yes 1,5,6,7 yes yes no
Qwerty sucks giga OSX, X11, Windows none no no 10 no no no

(*) Main goals of layouts

  1. Low same finger use
  2. Avoid adjacent finger use
  3. Rolls
  4. Prefer inward rolls over outward rolls
  5. High use of home row
  6. Avoid row jumps
  7. Finger load balance
  8. High hand alternation
  9. Resemblance to Qwerty
  10. Avoid typing machine jams

Software and real life use

Software: AdNW consists of a layout and several software implementations:

  • drivers & scripts for using the layout on Windows, Linux, OSX and *BSD; with versions for standard keyboards and for ergonomic keyboards (Ergodox, Maltron and others). This includes AutoHotKey files for Windows.
  • typing training software
  • the Optimizer: an algorithm (written in C++11, offered as source code) for calculating a custom AdNW-layout.

Userbase: AdNW is not (yet?) a massively adopted layout. The (German language) Google user group has around 100 participants, mostly from Germany. The website is in German too. However, the manual of the optimizer is in English, and participants on the discussion group say they are very open to questions in English.

Support: the developers of the software are active on the discussion forum and answer support questions.

Development: AdNW has been tested, alternative approaches have been tried, including multitudes of user designed AdNW versions. Usually, improvements in one area meant worsening in others. The developers claim that there is not much room for further overall improvement, but say they invite users to prove them wrong and to come up with new ideas.

BEAKL (2016)

BEAKL 9 on ortholinear board
BEAKL 8 Alt-Gr layer on ortholinear board
BEAKL 8 Numpad layer on ortholinear board
 
      40123 65987
      JHOUK GCRFZ
    - QIEAY DSTNB ;
      /,'.X WMLPV

http://shenafu.com/smf/index.php?topic=89.msg785#msg785

Advanced Keyboard Theory

BEAKL stands for Balanced Effortless Advanced Keyboard Layout. Its goals are as follows:

  • Balance the workload between each hand and between the fingers and keys based on the fingers' strength and agility and the keys' potential for rolls and speed.
  • Effortless means good rhythm in alternations and rolls, and getting a superior score in every keyboard metric, such as low distance, low penalties from same finger and same hands, etc. Possibly higher speed ceiling with less effort.
  • Advanced in the sense that it challenges traditional theories of typing. It asserts the following reasonings:
    • It does not advocate the so-called home row in touch typing. It vehemently avoids favoring the home pinky and the inside index key. The home pinky is extremely slow, weak, and uncomfortable to type. Instead, it strongly recommends the home block that consists of the ring, middle, and index fingers on the top, home, and bottom rows. These 9 keys form the core where the common letters (and sometimes punctuations) should be placed.
    • Related to the above, the pinky and inside index columns workload must be minimized. Each of these columns should have no more than 5% of the total key presses. Together, the two pinkies should do no more than 10% of the work; the two inside index columns together should not have more than 10% of the work. The latter is because the index home column already does a lot of work, so the inside column usage should be minimized.
    • The bottom ring and index keys are extremely fast and should not be avoided as other keyboard designers suggest. These can be sometimes up to 2x or faster than pressing the home pinky key.
    • Up and down finger movement is much faster, allows smoother rolls, and causes less strain than side-to-side movement (caused by pinky and inside index keys, and also by unnatural staggered keyboard design.)
    • The new theory can be summed up into an effort grid that rates the effort cost and roll potential of each key, pictured as follows. Lower values are better.

Keyboard effort grid.png

    • Recommend additional layers that are accessed by modifiers like shift, alt-gr, and numlock, so that fingers stay in the home block, rather than reaching all over the keyboard for numbers and symbols. This reduces time looking for symbols, speeds up writing code and math equations, provides an efficient numpad, and eliminates the need for tenkey numpad on the far right side of a full standard keyboard (thereby shrinking the size of the keyboard and allows mouse to be comfortably closer to the body). (See images in sidebar.)

Comparison with other layouts

Despite not putting the most common letters on the home row, the BEAKL layout still beats the competition in tests that heavily favor and expect common letters on the home row. This is accomplished with smart placement of common letters in the home block as described above. Thus achieving great combination of hand alternation and rolls with low same finger and same hand penalties. Due to its balanced nature, it is best to test this layout against other layouts using ergonomic keyboards designed with straight columns if possible.

The distance scores may be skewed due to the unnatural slant of standard keyboards. Most importantly, distance is not as good indicator as the actual time to hit a key. As claimed above, keys hit by the non-pinky fingers outside the home row are hit faster than the home pinky; even though generally the home pinky is thought of having a better distance score, while other faster keys may be penalized.

Modifications to AdNW Optimizer to accommodate new theory:

  • Effort scores for each key has been modified from AdNW's default to match the effort grid posted as shown above.
  • Heavily favor inward rolls compared to outward rolls.
  • Pinky target usage 2%, other fingers 15%.
  • Shift keys moved to thumb row (instead of outside pinkies.)
  • Opt defaults to 32 keys by adding 2 keys to the right side. So added -_ and /? and restricted them to outside the main 30 key block.

Results:

  • Beats everything else in AdNW optimizer with my customizations by far. MTGAP is closest but still a huge margin. Everything else is far behind.
  • Can compete head-to-head with MTGAP on other comparison tests using their own metric and effort values.

Layouts for specialised keyboards

Malt (1977)

KB Maltron.svg.png

Developed in conjunction with the two-handed Maltron physical keyboards.

The letter E is operated by one thumb. It is the most common letter in English, occurring in sequence with almost every other letter. There is major potential in giving E independence from the other fingers' movements. In fact, QWERTY with that one change would go a long way. In an optimized layout like this one, the change also clears another spot on the home-row where E would have been placed.

Malt puts priority the home and top rows, relegating rare keys to the bottom.

News from 1977 - somewhat sexist, says the keyboard weighs "only three pounds", and reassures you that it can be plugged into electric typewriters. It also cost $850, which in 2016 dollars is $2873.

Arensito (2001)

Implemented on a Kinesis
Implemented on a standard keyboard.

QL P   FUDK
ARENB GSITO
ZW HJ VCYMX

http://www.pvv.org/~hakonhal/main.cgi/keyboard

Also known as the Hallingstad layout. Arensito uses some unique punctuation arrangements, so they are not displayed above. Visit the link or check pictures for details.

Made by Håkon Hallingstad to suit the Kinesis Contoured and two-handed Maltron, though there is an adaptation for usage on standard keyboards too.

This layout is named after its home-row. The design intentions are as follows:

  • Places the eight most used characters under your fingertips.
  • Is the layout that minimizes the probability that you use the same finger twice (in succession).
  • Is the layout that maximizes the probability for using neighbour fingers in succession (and keeps the probability of sequences like y-d or z-l diminishingly low). This lets the fingers strike diagraphs and trigraphs extremely fast.
  • Keep the workload off the pinkies. Both pinkies press a button about 40% less than the other fingers.

Some punctuation and programming symbols are placed under the AltGr layer, where they are closer to the fingers' home position.

Amuseum (2008)

",WHM GCU.-
SOENL DTAIR
'(YVK BPF)X

http://shenafu.com/layout.php

Note: The Q, Z and J letters are placed outside the main-30 key area, ending up in locations determined by the keyboard used – standard or Kinesis Contoured.

The parentheses () turn into brackets with Shift activated.

The quote " and apostrophe ' signs turn into colon and semicolon, respectively, when Shifted.

The goals of Amuseum layout is to :

  1. balance the workload between each hand ,
  2. put common punctuation in the main block, even if that means putting rare letters outside ,
  3. mirror the keys for both hands. Notice vowels are mirrored and balanced on left and right hands, same with comma and period, and parentheses.

Advanced Developer's Dvorak (2010)

How ADDvorak looks on a semi-standard keyboard
One way to implement the Advanced Developer's Dvorak ergonomically.

https://web.archive.org/web/20130512083048/http://www.andong.co.uk/dvorak/DDvorak.aspx?page=Advanced+DDvorak

Also known as ADDvorak.

Keyboards that can be used with this layout are: any Japanese keyboard, Kinesis Contoured, Kinesis Freestyle, Maltron, Data Hand.

Created by freelance developer Andrei Stanescu, this layout is something out of the ordinary. It is only a Dvorak layout in the sense that almost all the letters are operated by the same fingers as they are in Dvorak.

At the outset, this layout relies on the concept that modifier-combinations are preferred over faraway keys.

Towards this purpose, there are six modifier states on top of the normal state. There are no utilised keys at all beyond the main home-key cluster of 3x4 for each hand. At least two thumb-keys are required for each hand.

Workman (2010)

QDRWB JFUP;
ASHTG YNEOI
ZXMCV KL,./

http://www.workmanlayout.com/

Created by OJ Bucao, it's a layout particularly made for straight-column keyboards.

The design crux is that of keeping the fingers to their main 4 columns per hand and de-prioritising columns in the middle and to the sides, thus minimizing diagonal and lateral movement.

Should be very suitable for a Kinesis Contoured-type of keyboard – the author also uses this.

Custom layout-generators

CarPalx

http://mkweb.bcgsc.ca/carpalx/

Created by bioinformatician Martin Krzywinski, this program will take any corpus of text and generate an optimal layout for typing it out. Works for non-English languages as well. Requires programming expertise and preferably Linux to set up.

Last updated: Version 0.11 from Aug 2009 (checked March 2012).

Using carPalx

The default setup of carPalx assumes a completely standard staggered keyboard. The proficient programmer should be able to modify it to his desires, however. But (s)he will have to come up with his own values on position penalties, for instance.

MTGAP

Program releases: https://github.com/MTGandP/Typing

In development by Michael Dickens since 2008, and found on the Mathematical Multicore blog. Inspired by Capewell's and Klausler's efforts. This program should now supersede both the Capewell and Klausler optimisers.

Every once in a while, the latest prime layout is posted on the blog. These layouts have used the "MTGAP" moniker, as in MTGAP 2.0 and MTGAP 3.5, though lately go unnamed (2012).

The blog comments were for a few years the center of discussion about layout design.

What it does

The program has different modes to optimise for specifically the Kinesis Contoured, or regular keyboards. These modes include predetermined values like penalties on each key-location, but such things can be edited to the user's wishes.

Setup on Windows

Requires GNU GCC. This means either boot to an Unix system, or use Cygwin or MinGW.

Using MinGW: Download it and upon installation include the C Compiler and MSys options.

Open up MinGW Shell: you should get a command-line window. Navigate your way with [cd …] commands to the MTGAP main program folder (it's the same directory you can find something called makefiles). Run "command make". An executable file should appear in the folder.

Setup on OS X

  • Install XCode. Go into Preferences and install Command Line Tools.
  • Open the terminal, navigate your way with [cd …] commands to the MTGAP program folder.
  • Run "make" to build the program. Run it again every time you change the .c or .h files.

Using the MTGAP optimiser

You will usually only have to modify the algorithm.c, keyboard.c, tools.c, values.c, values.h files and the relevant ‑layoutstore.txt file to customise your layout – all depending on how much you need to customise.

Anytime a non-textfile is edited, you must compile the program again before you run it (with "make", as described in the Setup section above).

Though it may be tempting to tailor the key-locations' penalties to thine own tastes (in values.c), keep in mind that such things as overworking the pinky-finger are already taken into account by other algorithms. The case is the same for the middle-finger's relative potency with heavy workload. Position penalties should only represent how hard the key actually is to reach.

AdNW Optimizer

http://adnw.de/index.php?n=Main.Downloads See there under Entwicklungsprogramme. Optimierer als C++-Quelltext, Häufigkeitstabellen für Deutsch und Englisch, Anleitung.

The AdNW Optimizer is a C++11 algorithm for calculating a custom AdNW-layout. AdNW is a Dvorak-type layout that values high hand alternation. The algorithm uses bigrams (2 letter combinations) or trigrams (three letter combinations), according to the users preference.

The download includes an English language manual. The algorithm is offered as source code, meaning the user should compile it before he can run it. Instructions are in the manual. Some user settings can be changed in the code before compiling, others can be changed at runtime. The program includes wordlists for English and German; input of other text corpi (prose, code) is possible. At runtime, users may optimize for mixes of languages. The optimizer can also be used to analyse and compare other layouts (Dvorak, Qwerty, Carplax, Colemak etc.) or to see what layout scores best for what language.

Outputs of the program include layouts, statistics and graphs.

Michael Capewell's Keyboard Evolve

Project appears inactive as of 2012[1] and has been since 2005[2]. Inspired by Klausler's evolver and supersedes it.

There are tools associated with this project that may prove useful to the budding layout-designer – a text-analyzer and typing tutor.

Peter M. Klausler's evolver

All links in circulation (http://www.visi.com/~pmk/evolved.html[3] & http://klausler.com/evolved.html[4]) appear to be dead. Capewell has presented Klausler's final layout for us though.

The Klausler layout (2002)

K,UYP WLMFC
OAEID RNTHS
Q.';Z XVGBJ

This layout is also known as the K,UYP layout by some. Klausler appears to have been using a Kinesis or the like, because he does not account for a standard staggered keyboard in the metrics.

Quoting Capewell on the calculations involved:

The score of a layout was given using the following criteria (paraphrased):

  • Every position has an assigned cost that's looked up from a table:
    • 53334 43335 (top row)
    • 10002 20001 (home row)
    • 64447 74446 (bottom row)
  • Using the same finger twice in succession on distinct letters costs 10 units.
  • When two keys in a row are struck with the same hand, it costs two units if they're on different rows or on the bottom row, and one unit if they're not adjacent.
  • If three or more keys are hit in succession by the same hand, one unit is charged for each key after the second.

References

  1. Last version update on KeyboardEvolve "Sep 19 2005", last checked 2012-03-04
  2. Michael Dickens, 2008-03-07, "He is supposedly is not finished, but it looks like he was doing this in 2005, so I think it’s as far as he’s going to get.", retrieved 2011-12-17
  3. Link found on this page, retrieved 2011-12-17
  4. Link found on The Big List, retrieved 2011-12-17