Welcome, Guest. Please login or register.
Did you miss your activation email?
2018-Oct-15 09:51

Login with username, password and session length

Recent

Shoutbox

Den:
Oct. 14 2018 - 5:03am
finally leveler cards is possible on card creator
Den:
Oct. 11 2018 - 12:43am
Vogon Full Art are so awesome
oohrah ;D
Den:
Oct. 09 2018 - 2:15am
KARA are Pretty Girls ... yeah yeah
Den:
Oct. 05 2018 - 2:02pm
Friday. get ready for more Last Man Standing  ;D
Den:
Oct. 03 2018 - 9:28pm
new card frames: Vogon HD, allows much more room for texts, incl. name, type, and esp. rules
Den:
Oct. 02 2018 - 7:48am
more card creator changes
Den:
Sep. 28 2018 - 5:08pm
Jim Morrison = Mr. Mojo Risin'
Den:
Sep. 19 2018 - 4:15pm
twit to image works again

Author Topic: Balanced Keyboard Layout  (Read 168896 times)

Den

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 1189
  • Selina is my Superstar
    • View Profile
    • Amuseum
Re: Balanced Keyboard Layout
« Reply #150 on: 2016-Sep-14 16:14 »
Mmmm...

https://geekhack.org/index.php?topic=73978.0

seems overboard and doesn't look that comfortable. thumbs shouldn't fan out that much. better to move whole arm and hand to hit farther keys.

the pinky is way too low, causing too much bend. it's not the natural at-rest position, and would be a pain to hit the lower pinky key. remember how we said we like stretches, not bends.
Support me on Patreon

I saw. I conquered. I came.

Den

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 1189
  • Selina is my Superstar
    • View Profile
    • Amuseum
Re: Balanced Keyboard Layout
« Reply #151 on: 2016-Sep-14 16:42 »
I've actually been thinking about why we are limited to 3 main rows (instead of 4 ... ) ... I suppose the touch typists think their fingers have to rest on some keys, instead of, for example, being between two rows ....

4 rows can be great if your typing style is moving around a lot, like Sean Wrona. Moving the whole arm is usually faster and less straining than stretching and bending only the fingers. But for traditional typists, extra rows doesn't seem conducive for the nice rolling action we are accustomed to.

Quote
I looked at the Arensito page, very old, but his initial design uses the top 3 rows rather than the bottom 3 ... in truth he requires a non-standard keyboard for his layout. I may revisit some of his ideas when I get around to my own design (once I get a logical layout that I'm happy with).
Arensito: http://www.pvv.org/~hakonhal/main.cgi/keyboard

His reason to move the letters up is so he can put all the special keys on the bottom row, which would be hit with the thumbs. Essentially mimicking the Kinesis and other keyboards that have a thumb cluster.

Den

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 1189
  • Selina is my Superstar
    • View Profile
    • Amuseum
Re: Balanced Keyboard Layout
« Reply #152 on: 2016-Sep-14 19:09 »
Have you tried altering "our" layouts to have Alt-Gr layer like Arensito? and compare stats.

Den

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 1189
  • Selina is my Superstar
    • View Profile
    • Amuseum
Re: Balanced Keyboard Layout
« Reply #153 on: 2016-Sep-15 00:48 »
To improve Stretch, consider if we remove the far top index and pinky finger and replace with the awesome under-index bend and side-pinky.

Code: [Select]
BEAKL Stretch Numdown

  "'x   wmb   
 qyouk gdlnv     
jhiea, cstrpz   
    .   f   


BEAKL Stretch Numup

  "'.   wmb   
 jyoux gdlnv   
qhiea, cstrpz
    k   f     

where the dot/K and F are now under the home index, and J/Q and Z are outside the home pinkies. The top index and pinkies are unassigned.

The effort has been improved by 15.% from the previous Stretch layout.

The most bottom row for Numdown contains the numerals. And naturally the dot/period is right beside them. In Numup, the number row stays at the top. This means this layout would require five rows, but the standard keyboard only has four. So Numup can only be used on keyboards like Kinesis, Ergodox, etc.
« Last Edit: 2016-Sep-15 03:58 by Den »

iandoug

  • Hero Member
  • *****
  • Posts: 880
    • View Profile
    • Keyboard Design
Re: Balanced Keyboard Layout
« Reply #154 on: 2016-Sep-15 02:37 »
Have you tried altering "our" layouts to have Alt-Gr layer like Arensito? and compare stats.

I tried with < > and wasn't sure of the effect... will try a better test later. Busy capturing the Ergo layout results now, up to around MTGAP Thumbshift which is so far the best. Tried comparing some of my Ergo layouts from earlier and they failed to beat it, but I have so many I'll have to do a systematic test and see which are good... and also try mapping something like current champ on ANSI onto Ergo layout... and the try tricks like putting H on thumb....

Den

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 1189
  • Selina is my Superstar
    • View Profile
    • Amuseum
Re: Balanced Keyboard Layout
« Reply #155 on: 2016-Sep-15 03:42 »
Comparing Arensito on standard layout and BEAKL on Ergo layouts, they have really close scores for code. Observations on why Arensito scores so highly on code.

1. The shift and alt-gr keys break up the same-finger and same-hand. So using more thumb-pressed shift and alt-gr give better scores.

2. Arensito has really low distance scores that are comparative to ergo layouts. This can be attributed to its mimicking of the thumb cluster as we noted earlier. Thereby lessening penalties from other fingers, particularly the pinkies.

3. Arensito has good locations for all the different types of brackets. Hitting them with alt-gr is still better than putting them on outside the pinkies, where they incur high penalties. Hence why it can get a good score on distance.

4. Arensito doesn't score so well on prose, as we are aware, but excellent on code. So we should try to steal its alt-gr trick for punctuation on a great prose layout, such as BEAKL and your alterations.
« Last Edit: 2016-Sep-15 03:51 by Den »

iandoug

  • Hero Member
  • *****
  • Posts: 880
    • View Profile
    • Keyboard Design
Re: Balanced Keyboard Layout
« Reply #156 on: 2016-Sep-15 04:06 »
3. Arensito has good locations for all the different types of brackets. Hitting them with alt-gr is still better than putting them on outside the pinkies, where they incur high penalties. Hence why it can get a good score on distance.

4. Arensito doesn't score so well on prose, as we are aware, but excellent on code. So we should try to steal its alt-gr trick for punctuation on a great prose layout, such as BEAKL and your alterations.

Mmm. Interesting idea. On the one hand it disrupts my plans for AltGr on my own keyboard design ... I must investigate the difference between "all brackets unshifted" vs "put them in easy-to-reach on AltGr". Also I currently have swapped AltGr with RShift, which may need a rethink.

Don't like things outside on the pinkies. But the keys are there and have to be used for something. On my "brackets" mod of current best layout, I've moved += (in that order) to where I have <> on top left. Don't like that they're both on same finger, but putting them that way is arguably easier to type than current Shift-+ = to get the common += in programs. I don't use my pinkies for those keys anyway as it's too short, I use ring finger. I should post a pic of my hands :-)

Gonna be busy till the weekend, so will see about those experiments then.

iandoug

  • Hero Member
  • *****
  • Posts: 880
    • View Profile
    • Keyboard Design
Re: Balanced Keyboard Layout
« Reply #157 on: 2016-Sep-15 10:46 »
Wonder how the community would feel about me adding this layout to the mix.... cry foul? But consider Arensito then....

http://patorjk.com/keyboard-layout-analyzer/#/load/xl5VRFRt

Does best with English not coding.



iandoug

  • Hero Member
  • *****
  • Posts: 880
    • View Profile
    • Keyboard Design
Re: Balanced Keyboard Layout
« Reply #158 on: 2016-Sep-16 14:17 »
Okay have uploaded the Ergo layouts, as well as new tests and interface updates.

Known issue: if you take too many results, it scrolls off the right of your screen, and no horizontal scrollbar. Not sure if browser bug or 'feature' of Semantic UI.
Will have to find a workaround.

BEAKL4 Ergo does quite well.

MTGap does well at prose, but not so well at other things.

Have not tested all the possibly permutations of inputs, so some combos might break or get MySQL errors/whatever ...please advise any problems :-)

I think I am finished adding tests now. Have other aspects to do/update/etc.

Cheers, Ian

iandoug

  • Hero Member
  • *****
  • Posts: 880
    • View Profile
    • Keyboard Design
Re: Balanced Keyboard Layout
« Reply #159 on: 2016-Sep-16 15:36 »
Here are conglomerate of BEAKL 1, 3, 4, 4 Ergo, and Stretch. Once you set home fingers to the bottom row, it scores very well. Almost as good as BEAKL 3.

Code: [Select]
BEAKL Opted4 Ergo
73.55

Would you prefer 74.74 on Alice? :)

Just a quick re-arrange based on current top ANSI layout.
http://patorjk.com/keyboard-layout-analyzer/#/load/W3X4Cswn

All three metrics on all three tests on Patrick dramatically better. Will check other inputs later.

Den

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 1189
  • Selina is my Superstar
    • View Profile
    • Amuseum
Re: Balanced Keyboard Layout
« Reply #160 on: 2016-Sep-16 16:11 »
Quote
Don't like things outside on the pinkies. But the keys are there and have to be used for something.

They're much better spots for Ctrl, Alt, etc. than at the bottom. Your pinkies don't have to stretch as much when doing one-handed combinations.

Den

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 1189
  • Selina is my Superstar
    • View Profile
    • Amuseum
Re: Balanced Keyboard Layout
« Reply #161 on: 2016-Sep-17 02:43 »
Beat Arensito at code at its own game:
http://patorjk.com/keyboard-layout-analyzer/#/load/qr7jx9Dm

Check out BEAKL Opted4 Ergo Alt, scoring upper 60's for all sorts of code (including javascript, html, java, json).
« Last Edit: 2016-Sep-17 02:59 by Den »

iandoug

  • Hero Member
  • *****
  • Posts: 880
    • View Profile
    • Keyboard Design
Re: Balanced Keyboard Layout
« Reply #162 on: 2016-Sep-17 03:21 »
Beat Arensito at code at its own game:
http://patorjk.com/keyboard-layout-analyzer/#/load/qr7jx9Dm

Check out BEAKL Opted4 Ergo Alt, scoring upper 60's for all sorts of code (including javascript, html, java, json).

Way cool :-)
I checked the heat map page, it does look like KLA is indeed counting the keys on AltGr.... I see you've even put the digits on AltGr.... interesting idea... means it frees up the awkward locations for REALLY UNUSED characters.

I'll have to ponder your ideas a bit later... promised a client a job would be done by Monday so that's got to take priority... :-)

BTW while compiling all my stats, I noticed another layout which also had double-decker digits ... so what I thought was a totally new idea is not actually...  nothing new under the sun... :-)

Den

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 1189
  • Selina is my Superstar
    • View Profile
    • Amuseum
Re: Balanced Keyboard Layout
« Reply #163 on: 2016-Sep-17 05:17 »
Using Alt-Gr to game the scores is not necessarily more ergonomic, though.  there's a lot of value to hit special characters without modifiers. shortcuts in apps, especially games that need quick, simple key presses, especially the numbers. I like to use keys at the edge of the keyboard to control playback in VLC.

iandoug

  • Hero Member
  • *****
  • Posts: 880
    • View Profile
    • Keyboard Design
Re: Balanced Keyboard Layout
« Reply #164 on: 2016-Sep-17 06:36 »
Yeah, I guess we'll have to try using a keyboard configured that way. I also prefer frequently used things to be unshifted and easy to hit. Which is one of the reasons why my keyboard has so many keys... :-)

Anyway I knocked together a quick and dirty frequency counter, here's the results for the programming inputs. Set below it includes my PHP and Perl programs, which I can't post publicly.

Also = is used almost as much (which I nearly typed as almuch now...) as the quotes ... this means my current location up left is way unoptimal.... (yes spellchecker I know it's not a word...)
Includes some high-ASCII characters, I looked them up, mostly European plus dagger and degree and per-mil.

  : &#x20; : 49922
. : &#x2e; : 10111
, : &#x2c; : 7283
= : &#x3d; : 7150
" : &#x22; : 7093
) : &#x29; : 6754
( : &#x28; : 6747

 : &#x0a; : 5907
; : &#x3b; : 5370
- : &#x2d; : 5090
: : &#x3a; : 4184
0 : &#x30; : 3642
_ : &#x5f; : 3621
{ : &#x7b; : 3339
} : &#x7d; : 3335
1 : &#x31; : 3063
< : &#x3c; : 3004
> : &#x3e; : 2970
/ : &#x2f; : 2085
2 : &#x32; : 1635
& : &#x26; : 1547
[ : &#x5b; : 1380
* : &#x2a; : 1377
] : &#x5d; : 1368
+ : &#x2b; : 1323
3 : &#x33; : 1203
4 : &#x34; : 1077
| : &#x7c; : 1021
         : &#x09; : 986
# : &#x23; : 956
' : &#x27; : 955
! : &#x21; : 954
5 : &#x35; : 932
8 : &#x38; : 753
6 : &#x36; : 740
9 : &#x39; : 638
7 : &#x37; : 636
? : &#x3f; : 532
$ : &#x24; : 523
% : &#x25; : 296
\ : &#x5c; : 286
� : &#xe2; : 208
~ : &#x7e; : 187
� : &#x94; : 153
� : &#x80; : 148
@ : &#x40; : 147
^ : &#x5e; : 64
� : &#x86; : 44
� : &#x92; : 33
` : &#x60; : 22
� : &#x89; : 9
� : &#x91; : 5
� : &#xa4; : 5
� : &#xa0; : 4
� : &#xc2; : 4
� : &#xb7; : 3
� : &#x93; : 3
� : &#x90; : 3
� : &#x9c; : 2
� : &#xef; : 2
� : &#xb0; : 1
� : &#xab; : 1
� : &#x96; : 1
� : &#xba; : 1


-----------------------------------
Include php and pl program with above. pl program generates html/css/php/javascript.
As expected $ and # frequency go up.

  : &#x20; : 133436
" : &#x22; : 23622

 : &#x0a; : 16514
, : &#x2c; : 15643
# : &#x23; : 15466
. : &#x2e; : 12317
= : &#x3d; : 10967
) : &#x29; : 10624
( : &#x28; : 10618
; : &#x3b; : 10488
$ : &#x24; : 8742
0 : &#x30; : 6467
- : &#x2d; : 6434
         : &#x09; : 6406
/ : &#x2f; : 5603
1 : &#x31; : 5569
} : &#x7d; : 5146
{ : &#x7b; : 5144
> : &#x3e; : 5103
: : &#x3a; : 5031
< : &#x3c; : 4984
_ : &#x5f; : 4587
2 : &#x32; : 3496
3 : &#x33; : 2615
' : &#x27; : 2323
4 : &#x34; : 2248
~ : &#x7e; : 2146
& : &#x26; : 2145
6 : &#x36; : 2103
8 : &#x38; : 1883
5 : &#x35; : 1866
\ : &#x5c; : 1673
* : &#x2a; : 1636
+ : &#x2b; : 1609
7 : &#x37; : 1599
[ : &#x5b; : 1567
] : &#x5d; : 1536
! : &#x21; : 1281
9 : &#x39; : 1173
| : &#x7c; : 1164
? : &#x3f; : 658
@ : &#x40; : 602
% : &#x25; : 541
� : &#xe2; : 210
^ : &#x5e; : 163
� : &#x94; : 153
� : &#x80; : 150
� : &#x86; : 44
� : &#x92; : 33
` : &#x60; : 22
� : &#x89; : 9
� : &#x91; : 5
� : &#xa4; : 5
� : &#xc2; : 4
� : &#xa0; : 4
� : &#x90; : 3
� : &#x93; : 3
� : &#xb7; : 3
� : &#x99; : 2
� : &#xef; : 2
� : &#x9c; : 2
� : &#x96; : 1
� : &#xb0; : 1
� : &#xab; : 1
� : &#xba; : 1
« Last Edit: 2016-Sep-17 08:44 by iandoug »

iandoug

  • Hero Member
  • *****
  • Posts: 880
    • View Profile
    • Keyboard Design
Re: Balanced Keyboard Layout
« Reply #165 on: 2016-Sep-17 12:12 »
Input: 500 ISO dates. I was impressed with how well Beakl4 mod Ian did against the usual contenders (MTGAP, Colemak Programmer, etc), until I added Arensito ... then adding your latest alt ergo made me feel a little better... :-)

Rank         Layout         Score
#1 BEAKL Opted4 Ergo Alt 57.94
#2 Arensito fixed 57.11
#3 BEAKL 4 Mod Ian (best one) 47.20
#4 MTGAP 16.36
#5 Personalized 16.36
#6 Programmer Colemak 15.53

It's kinda weird that scores can differ so vastly for what is basically numeric input ....


iandoug

  • Hero Member
  • *****
  • Posts: 880
    • View Profile
    • Keyboard Design
Re: Balanced Keyboard Layout
« Reply #166 on: 2016-Sep-17 15:19 »
Mmm. 500 ISO DateTimes with random timezone offsets. Arensito tricks not always helpful...
eg: 1994-09-19T23:02:50+20:00.   Dates from 1970 to 31 Dec 2030.

Rank      Layout                Score
#1       BEAKL Opted4 Ergo 47.71
#2      BEAKL 4 Mod Ian    45.08
#3      BEAKL Opted4 Ergo Alt 44.01
#4      Arensito fixed 42.19
#5      Simplified Dvorak 30.70
#6      Personalized        21.38

iandoug

  • Hero Member
  • *****
  • Posts: 880
    • View Profile
    • Keyboard Design
Re: Balanced Keyboard Layout
« Reply #167 on: 2016-Sep-19 05:22 »
Started playing around with AltGr ... starting from Beakl 4 Mod Ian best.

http://patorjk.com/keyboard-layout-analyzer/#/load/wmLflcsd

Need to work on the digits more... tried moving K and V to the blank spot but scores went down on KLE.

iandoug

  • Hero Member
  • *****
  • Posts: 880
    • View Profile
    • Keyboard Design
Re: Balanced Keyboard Layout
« Reply #168 on: 2016-Sep-19 07:22 »
Well that was depressing ... had a nice variant catching up to yours and Konqueror goes and crashes on me ... and no recent saved version. :-(

Sigh. :-)

Now need to try and remember what I did ...

iandoug

  • Hero Member
  • *****
  • Posts: 880
    • View Profile
    • Keyboard Design
Re: Balanced Keyboard Layout
« Reply #169 on: 2016-Sep-19 10:11 »
Wins most programming but not all.

http://patorjk.com/keyboard-layout-analyzer/#/load/ntLmFD3r

SameFinger metrics good, Distance not so good. First had to overcome the +/- 3 point loss from swapping Shift and AltGr. So in the end, pretty good.
73,78 on Alice...

Not sure how practical this whole AltGr move is ... might work better on a non-ANSI layout. For example I often have to add a comma to the end of a bunch of lines in a text editor. So my right hand is tied up with the arrow-down and End keys, it can't also be pressing AltGr so the left can hit the comma on AltGr whatever. Think I did try with comma and period as Primary, but think I got better scores putting them as alt on home row ... will fiddle more later.

Also you hand is unnaturally contorted keeping the thumb on AltGr... which is why a different layout will be better. Suppose keyboards need a shorter space bar ...

iandoug

  • Hero Member
  • *****
  • Posts: 880
    • View Profile
    • Keyboard Design
Re: Balanced Keyboard Layout
« Reply #170 on: 2016-Sep-19 18:03 »
Ignore message layout about ... missing the colon.

iandoug

  • Hero Member
  • *****
  • Posts: 880
    • View Profile
    • Keyboard Design
Arensito is annoying...
« Reply #171 on: 2016-Sep-20 13:18 »
So after getting a layout that wipes the floor with Arensito I add another test:

500 dates in German format
dd.mm.yyyy

and it promptly beats me again.

Good going for a layout developed so long ago....

Sigh :-)

iandoug

  • Hero Member
  • *****
  • Posts: 880
    • View Profile
    • Keyboard Design
Re: Balanced Keyboard Layout
« Reply #172 on: 2016-Sep-20 14:44 »
Part of Arensito's "secret" is that AltGr is on the spacebar, and the only shift key directly above it. So the distance that the left thumb has to move is one key.
By comparison I have AltGr and RShift on right thumb next to each other , but the distance from centre of AltGr to RShift is 1.5 keys, because AltGr is wider.

All those extra 0.5 keyshifts add up, leading to a higher distance, and lower score, I imagine. I am loathe to copy that idea... at the moment I'm sitting with a layout that has only 4 rows instead of 5 ... numrow is completely blank and free for esoteric uses.... :-)

iandoug

  • Hero Member
  • *****
  • Posts: 880
    • View Profile
    • Keyboard Design
Arensito
« Reply #173 on: 2016-Sep-23 10:18 »
Hi

http://patorjk.com/keyboard-layout-analyzer/#/load/j0VSsQLk

What's so special about his number layout? Or is it just that ours sucks in comparison?
1000 digits of π

iandoug

  • Hero Member
  • *****
  • Posts: 880
    • View Profile
    • Keyboard Design
Numpads
« Reply #174 on: 2016-Sep-23 17:00 »
Mmmm.

So does this mean that numpads, while looking like a good idea, are not actually?

http://patorjk.com/keyboard-layout-analyzer/#/load/Hlmsrmrm

I really don't like the direction this research is sending me ... more and more towards keyboard.io ... I realised that I could fit the entire keyboard onto 4 rows (including bottom modifiers) by using AltGr+shift (or another dedicated modifier).

Which is of course completely the opposite direction I was going in when I started designing my keyboard ... but how must I argue with the numbers? :-)
« Last Edit: 2016-Sep-24 06:47 by iandoug »

iandoug

  • Hero Member
  • *****
  • Posts: 880
    • View Profile
    • Keyboard Design
Re: Balanced Keyboard Layout
« Reply #175 on: 2016-Sep-23 17:03 »
I suppose numpads are good when you're keeping your place on a page with the left hand and entering lots of numbers with the right ... else you might need three hands....

Read a joke today that if aliens landed here in  10000 years time and tried to figure out how we looked from the keyboards lying around, they'd conclude we had ten tentacles coming out of our chests....

iandoug

  • Hero Member
  • *****
  • Posts: 880
    • View Profile
    • Keyboard Design
backspace/delete
« Reply #176 on: 2016-Sep-24 08:39 »
Thoughts on putting Backspace on AltGr-space?

And maybe delete on AltGr-shift-space?

https://autohotkey.com/board/topic/109966-ergonomic-coding-script-for-me-atleast/

Thanks, Ian

Den

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 1189
  • Selina is my Superstar
    • View Profile
    • Amuseum
Re: Arensito
« Reply #177 on: 2016-Sep-24 16:38 »
Hi

http://patorjk.com/keyboard-layout-analyzer/#/load/j0VSsQLk

What's so special about his number layout? Or is it just that ours sucks in comparison?
1000 digits of π

My Ergo Alt got 49.47, compared to Arensito 46.77. On standard keyboard, 4 points behind Arensito. If I put 2 on the index home, they both gain 3 points.
http://patorjk.com/keyboard-layout-analyzer/#/load/k7l8Zl51
Code: [Select]
BEAKL Opted4 Ergo Alt
49.47 (52.17)

Arensito fixed
46.77

BEAKL Opted4 Alt
45.48

Personalized
40.22

BEAKL 4 Mod Ian AltGr KLE68.79! check
38.65

BEAKL Opted4 Ergo Mod Ian 1
38.56

« Last Edit: 2016-Sep-24 16:55 by Den »

Den

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 1189
  • Selina is my Superstar
    • View Profile
    • Amuseum
Re: Balanced Keyboard Layout
« Reply #178 on: 2016-Sep-24 17:09 »
What we can learn from numpads:

1. Numpad are 1-handed only. This is very stressful on that hand, not always good rolls, and high same-finger.

2. It uses only the 3 strong fingers (and 0 on the thumb) for the digits, but no pinky. This is the exact same thing I've been advocating in BEAKL: No common letter/number on the pinky.

3. Another familiar concept is the stretching. Since 123 are the most common numbers, that is our home row. That means we stretch to reach the other numbers.

That said, I can envision a different (better?) design for 1-hand numpad based on BEAKL philosophy and effort grid.
« Last Edit: 2016-Sep-24 17:31 by Den »

Den

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 1189
  • Selina is my Superstar
    • View Profile
    • Amuseum
Re: backspace/delete
« Reply #179 on: 2016-Sep-24 17:14 »
Thoughts on putting Backspace on AltGr-space?

And maybe delete on AltGr-shift-space?

https://autohotkey.com/board/topic/109966-ergonomic-coding-script-for-me-atleast/

Thanks, Ian

I much prefer Backspace on its own thumb key, and the Kinesis makes it possible. It doesn't affect these test scores, anyway. (I've heard of some programs that simulate typing mistakes, and thus include random backspace into their tests.)


iandoug

  • Hero Member
  • *****
  • Posts: 880
    • View Profile
    • Keyboard Design
Re: Balanced Keyboard Layout
« Reply #180 on: 2016-Sep-24 17:29 »
2. It uses only the 3 strong fingers (and 0 on the thumb) for the digits, but no pinky. This is the exact same thing I've been advocating in BEAKL: No common letter/number on the pinky.

Enter/return is on pinky.

Curiously, if I remove the need for the enter key, and paste pi all as one string, the numpad scores goes DOWN (which is not what I was expecting :-) )
http://patorjk.com/keyboard-layout-analyzer/#/load/8crzN68C

And you do better...

http://patorjk.com/keyboard-layout-analyzer/#/load/VJz26H4f

I'm supposed to be working ...  I need to redo my AltGr, spent most of the day fixing my site. Yeah, that meant fighting with Javascript to get the results into nicely scrollable tables that also sort properly. Then had to deal with the site management software having mangled a few things as well ;)

So AltGr will have to wait for another day....

iandoug

  • Hero Member
  • *****
  • Posts: 880
    • View Profile
    • Keyboard Design
Re: Balanced Keyboard Layout
« Reply #181 on: 2016-Sep-24 17:33 »
and so we stumble onto another of Arensito's secrets .... the Enter key way up there... but REALLY close to the pinky .... so any tests with lots of Enters gives it an advantage over ANY other layout that leaves Enter in default position.

I've moved it in my AltGr, but not as short a distance as he has... I suppose I need to put it on a single key that the distance calculations become more favourable....

iandoug

  • Hero Member
  • *****
  • Posts: 880
    • View Profile
    • Keyboard Design
Re: Balanced Keyboard Layout
« Reply #182 on: 2016-Sep-24 17:40 »
mmm something like this makes rather a radical difference

http://patorjk.com/keyboard-layout-analyzer/#/load/23xmLtmc

see BEAKL 4 Mod Ian AltGr KLE68.79! check layout. Almost as good as an Ergo layout...

iandoug

  • Hero Member
  • *****
  • Posts: 880
    • View Profile
    • Keyboard Design
Re: backspace/delete
« Reply #183 on: 2016-Sep-24 17:43 »
I much prefer Backspace on its own thumb key,

If your thumb homes are Space and AltGr, then backspace becomes a no-mover...

I'll try it sometime... in reality my thumbs hover over the board (like all the other fingers not actually pressing anything...) so it might turn out to be more effort... ;-)

Den

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 1189
  • Selina is my Superstar
    • View Profile
    • Amuseum
Re: Balanced Keyboard Layout
« Reply #184 on: 2016-Sep-24 19:54 »
Quote
Enter/return is on pinky.

Enter usage varies wildly depending on type of entry.

Ideally Enter would be accessed by the thumb.

Preliminary optimized layout looks like this:

Code: [Select]
BEAKL OptedN Nums

 543 
70216
 8 9
.

There are 3 rows and 5 columns, and 2 more thumb keys. That is exactly the same number of keys on a standard numpad. I haven't included the math signs or Enter yet.

Visually:


With signs:
Code: [Select]
BEAKL OptedN Full

754=* 
31/2+
908-6
.     

Visually:


This works great with dates, too, both with slashes or dashes.

What if we needed more signs?

Code: [Select]
BEAKL OptedN Ext

^+-=*
%62.#
9315:
784$/
0

Visually:
« Last Edit: 2016-Sep-24 21:39 by Den »

Den

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 1189
  • Selina is my Superstar
    • View Profile
    • Amuseum
Re: Balanced Keyboard Layout
« Reply #185 on: 2016-Sep-26 15:22 »
My next breakthrough is to expand the pinkies in both directions. This not only improves the scores on prose, but can almost match Arensito on code without resorting to cumbersome AltGr.

This next evolution I dub BEAKL Opted36 because it 6 columns x 3 rows x 2 hands, for total 36 keys.
Code: [Select]
BEAKL Opted36

{jyou( wdmn)z
xhiea. gstrpk
}'"/,; fclbvq

see image


in KLA :http://patorjk.com/keyboard-layout-analyzer/#/load/bnKXn59Z

Code: [Select]
Results

Alice
72.74

PPTT
31.

Google source
53.00
« Last Edit: 2016-Sep-26 15:36 by Den »

iandoug

  • Hero Member
  • *****
  • Posts: 880
    • View Profile
    • Keyboard Design
Re: Balanced Keyboard Layout
« Reply #186 on: 2016-Sep-26 15:36 »
Mmm interesting.... I'm actually going in the other direction and moving towards the centre, but that requires considerable use of AltGr, which as Arensito pointed out way back when, in not very comforable on ANSI layout... need custom design.

I found scores improve if you move Enter to where ''/" is, but that's only doable if you can shuffle keys around via AltGr.

Got very upset this morning, had a new layout that got new high scores on English prose and then I lost it ... thought I had it saved but didn't, it seems. So I tried again and got one that was actually better at English but worse on the technicals.
It's layout 1 here (the others are the variants I was fiddling with).
http://patorjk.com/keyboard-layout-analyzer/#/load/GrGK0QRX

Good scores, low distances, samefinger not the best but not too bad either.
The / is certainly a problem, it needs to move.

Wasn't going to share it until I had sorted out the technicals but then you popped along... :-)

I will contemplate your ideas above. I can see one issue, is that you are using WIDE keys which means Patrick's calculations for distance will hit you....

iandoug

  • Hero Member
  • *****
  • Posts: 880
    • View Profile
    • Keyboard Design
Re: Balanced Keyboard Layout
« Reply #187 on: 2016-Sep-26 15:39 »
The copyright sign and "not" sign are just to force KLA to put the legend at the bottom of the keycap instead of the top when there is not shifted character.

iandoug

  • Hero Member
  • *****
  • Posts: 880
    • View Profile
    • Keyboard Design
Re: Balanced Keyboard Layout
« Reply #188 on: 2016-Sep-26 19:20 »
 I wonder if 75 on Alice is doable on ANSI?.... getting close.

http://patorjk.com/keyboard-layout-analyzer/#/load/PVLp2rKm

Den

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 1189
  • Selina is my Superstar
    • View Profile
    • Amuseum
Re: Balanced Keyboard Layout
« Reply #189 on: 2016-Sep-26 20:28 »
Even better when you bring 1 and 0 down.

Code: [Select]
qjyou( wdmn0{
xhiea. gstrpv
}1"),: fclbk



http://patorjk.com/keyboard-layout-analyzer/#/load/wXXP350C
« Last Edit: 2016-Sep-27 05:48 by Den »

Den

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 1189
  • Selina is my Superstar
    • View Profile
    • Amuseum
Re: Balanced Keyboard Layout
« Reply #190 on: 2016-Sep-27 06:06 »
Center keys do seem to be effective. BEAKL 4 Ergo already uses them and does very well on code. The problem is Kinesis doesn't have those keys, so I can't try such a layout.


Den

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 1189
  • Selina is my Superstar
    • View Profile
    • Amuseum
Re: Balanced Keyboard Layout
« Reply #191 on: 2016-Sep-27 06:17 »
and so we stumble onto another of Arensito's secrets .... the Enter key way up there... but REALLY close to the pinky .... so any tests with lots of Enters gives it an advantage over ANY other layout that leaves Enter in default position.

I've moved it in my AltGr, but not as short a distance as he has... I suppose I need to put it on a single key that the distance calculations become more favourable....

the fact Arensito use pinky for Enter is huge faux-pas. that could be up to 20% for each pinky. ouch! maybe try AltGr + middle finger.

iandoug

  • Hero Member
  • *****
  • Posts: 880
    • View Profile
    • Keyboard Design
Re: Balanced Keyboard Layout
« Reply #192 on: 2016-Sep-27 06:44 »
the fact Arensito use pinky for Enter is huge faux-pas. that could be up to 20% for each pinky. ouch! maybe try AltGr + middle finger.

According to Michael's frequency tables, Enter is between w and b ... Mmm that's an idea.
What I was trying to say was that it's right down there in QXZ territory, not up in "should be on a thumb" territory. I think if entering data (numbers etc) then we use enter a lot, but if writing text we don't. If coding, then a lot more. I think Michael used formal text, informal text and code to get his frequencies, so given that code would have pushed the frequency higher, in normal text it will be very low... certainly in modern word processors, only for paragraphs not end of line.

YMMV :-)

Den

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 1189
  • Selina is my Superstar
    • View Profile
    • Amuseum
Re: Balanced Keyboard Layout
« Reply #193 on: 2016-Sep-27 08:19 »
According to Michael's frequency tables, Enter is between w and b ... Mmm that's an idea.
What I was trying to say was that it's right down there in QXZ territory, not up in "should be on a thumb" territory. I think if entering data (numbers etc) then we use enter a lot, but if writing text we don't. If coding, then a lot more. I think Michael used formal text, informal text and code to get his frequencies, so given that code would have pushed the frequency higher, in normal text it will be very low... certainly in modern word processors, only for paragraphs not end of line.

YMMV :-)

there's also a high chance of multiple Enters hit consecutively, and that's even worse for the pinky. that's how i probably hurt my right pinky on keyboard at work.

iandoug

  • Hero Member
  • *****
  • Posts: 880
    • View Profile
    • Keyboard Design
Re: Balanced Keyboard Layout
« Reply #194 on: 2016-Sep-27 09:06 »
FWIW my own design put the enter on the thumbs... I think we've subconsciously associated 'big key' == important, even if that's not true (and a hangover from typewriters).

http://www.keyboard-layout-editor.com/#/samples/pkb.json

Said design has now largely been chucked out the window, based on last few month's learning.. :-)

iandoug

  • Hero Member
  • *****
  • Posts: 880
    • View Profile
    • Keyboard Design
75 today...
« Reply #195 on: 2016-Sep-27 13:44 »
Well almost ... short with 1c. Maybe you can find the missing cent.

Alice, 74.99  on ANSI. Also a layout I did back in July that got top score on Alice.

ANSI layout is optimised for Alice ... punctuation in Alice is placed to get good score. Other punctuation not attended to. So probably not much good for other stuff, though should be reasonable for other English prose, depending if single or doublequoted.

http://patorjk.com/keyboard-layout-analyzer/#/load/KGdP8Zlj

Just tried it to see if it was possible to hit 75 (based on the assumption that 75 is 3/4 of the way to 100, meaning the keyboard is only 3/4 optimised.... which is depressing :-)  )

iandoug

  • Hero Member
  • *****
  • Posts: 880
    • View Profile
    • Keyboard Design
shift and alts
« Reply #196 on: 2016-Sep-27 13:56 »
People are used to using shift to access things on a key ... AltGr is just another shift with a silly name from the 70s.

The biggest problem we have at the moment is that the bureaucrats have declared ANSI and ISO as The Standards and everything else, no matter how much better, is treated with suspicion.

Based on current top-scoring layouts, we can make a much better keyboard with appropriately sized and spaced keys, free of legacy influences.

As I said before, it possible to put the rest of the ANSI 104/105 keys onto 4 rows (no numpad though, that's a duplication). However then there is no space for certain things already mapped to AltGr like Euro symbol, Euro letters ë ç ø etc on some National layouts. See AltGr page on Wikipedia  https://en.wikipedia.org/wiki/AltGr_key

Den

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 1189
  • Selina is my Superstar
    • View Profile
    • Amuseum
Re: shift and alts
« Reply #197 on: 2016-Sep-27 14:18 »
People are used to using shift to access things on a key ... AltGr is just another shift with a silly name from the 70s.

The biggest problem we have at the moment is that the bureaucrats have declared ANSI and ISO as The Standards and everything else, no matter how much better, is treated with suspicion.

Based on current top-scoring layouts, we can make a much better keyboard with appropriately sized and spaced keys, free of legacy influences.

As I said before, it possible to put the rest of the ANSI 104/105 keys onto 4 rows (no numpad though, that's a duplication). However then there is no space for certain things already mapped to AltGr like Euro symbol, Euro letters ë ç ø etc on some National layouts. See AltGr page on Wikipedia  https://en.wikipedia.org/wiki/AltGr_key
Simply propose a new modifier, the Punc key, that accesses a new layer. Actually that sounds like Func / Fn which is found on laptops and other keyboards for media functions. Could make that standard on all new keyboards henceforth.

iandoug

  • Hero Member
  • *****
  • Posts: 880
    • View Profile
    • Keyboard Design
Re: Balanced Keyboard Layout
« Reply #198 on: 2016-Sep-27 14:36 »
:-)

And we label it with an interrobang? Or two? :-)

9‽
‽‽
Or maybe Pilcrow for Punctuation ¶ but I suppose that could be confusing as some northern Euro layouts have that sign on already (to actually use it).

Or ⁂ would be most elegant..... and possibly confusing :-)
Stolen from http://mentalfloss.com/article/12710/13-little-known-punctuation-marks-we-should-be-using

Maybe just call it the NumPunc key .... :-)



iandoug

  • Hero Member
  • *****
  • Posts: 880
    • View Profile
    • Keyboard Design
Re: Balanced Keyboard Layout
« Reply #199 on: 2016-Sep-27 14:40 »
Center keys do seem to be effective. BEAKL 4 Ergo already uses them and does very well on code. The problem is Kinesis doesn't have those keys, so I can't try such a layout.

I didn't know what to do with those keys when I was fiddling with the ergo layouts ... their shape implies they're not for letters, but for modifiers/whatever, and it felt odd putting letters etc there. I did eventually dig up some 'official' key mappings for those layouts but haven't had time to play with them yet.

 

7 Guests, 0 Users