ASETNIOP or 6x2

[Craig Tyle]

ASETNIOP is a "chord-based" keyboard layout that uses the most frequent letter for each finger on a qwerty keyboard as its foundation.

I wonder how much efficiency could be achieved if one adopted this approach for an ordinary (i.e., non-chord) keyboard.  Such a keyboard (call it "6x2") would involve six swaps of keys -- E-D, T-F, N-J, I-K, O-L, and P-;.  You're just moving the most frequent letter in each case to the home row.  It would seem far easier to transition to such a layout, since there are no finger changes.  So, how efficient would it be compared to Minimak or Colemak - would you get more "bang for the buck" even if overall efficiency were a little lower?

Obviously, by definition, such a layout would not address left-right imbalance in qwerty, as no finger changes.  But one additional swap (call the resulting keyboard "6+1") could get you there: you could swap the (moved) E and I keys (yielding a home row of ASITGHNEOP) or the E and N keys (ASNTGHEIOP).  The former is pretty close to Minimak 12.  Both are even closer to qwpr, but avoid moving the E to the right pinky.  

I wonder (1) how much additional improvement you would get from this additional move (and, thus, whether it would be worth it), (2) which version of "6+1" would be better (I suspect the former, as both Colemak and Minimak 8 have the NEO combination), and (3) how close 6+1 would come to Minimak or Colemak in terms of overall efficiency.

Best regards. 

[Rob Neff]

Just from my experience, it doesn't seem to matter much if the key stays on the same finger or not. There's still going to be a learning curve for every key that changes, so might as well optimize where that key is (ie. left/right balance) where possible.  YMMV.  I wonder how to get objective data on learning curves for your 6x2 approach vs. minimak-type swap.

I'm using a 6-letter swap (minimak-4 plus 2 from minimak-8): E, D, T, J, N, K.  Given the information on the minimak page that should be about 75% of the theoretical efficiency gain of Dvorak, but with only 6 letters.  I've found the J/K swaps mess with vi, but I almost never use that editor anymore.  All my other shortcuts and such are untouched or easily adapted.

Rob

[Pit]

I know that the e-mail is loooong time ago, but nevertheless maybe someone is thinking the same. I also tried the 6x2 layout in a slight variation (ed, rf, tg, nj, ik, ol) and found that this will give some improvement, but will not be worthwhile IMO, because same-finger-repetition does not improve as well the imbalance of qwerty. So one _has_ to make some more serious changes. Your idea of 6x2+1 is a nice one IMO. I just tested that in one analyzer:

http://klanext.keyboard-design.com

The 6x2 layouts are better than qwerty, but worse than Minimak 4, so they will not make sense IMO. Would have been too nice :-)

Here the detailed results of some contenders (Minimaks is a layout I suggested in another thread -- a further developed, but more complicated Minimak 8 version). Those results are for English. For other languages the results will change (can be significant).

Rank

Layout

Board

+Effort

Overall Score

Distance

Same Finger

Same Hand

#1

adnw.en.ansi

standard

+0%

161.36

69.52

81.42

10.42

#2

k-u_tea_shrt-pnct.en.ansi

standard

+0%

162.16

70.29

80.41

11.46

#3

poqtea-qp.en.ansi

standard

+3%

166.28

70.53

83.77

11.98

#4

colemak.en.ansi

standard

+5%

168.80

66.61

88.27

13.93

#5

Minimaks

standard

+18%

190.69

73.62

100.31

16.76

#6

asset-corrected.en.ansi

standard

+19%

191.45

67.31

109.58

14.55

#7

minimak-12key.en.ansi

standard

+23%

198.41

71.43

110.22

16.76

#8

minimak-8key.en.ansi

standard

+26%

203.66

76.69

110.22

16.76

#9

minimak-4key.en.ansi

standard

+32%

213.18

86.20

110.22

16.76

#10

6x2+1

standard

+32%

213.30

74.40

122.78

16.12

#11

6x2

standard

+37%

220.69

73.96

131.97

14.76

#12

qwerty.en.ansi

standard

+52%

245.49

98.77

131.97

14.76

Op donderdag 23 maart 2017 om 15:04:58 UTC+1 schreef csty...@gmail.com:

[paper2222]

ay caramba

this is a bad analyzer

--
You received this message because you are subscribed to the Google Groups "Minimak" group.
To unsubscribe from this group and stop receiving emails from it, send an email to minimak-layou...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/minimak-layout/a8fbe3a3-0fde-4741-b728-621aeb4f5737n%40googlegroups.com.

[Pit]

I take the analyzers with a grain of salt. The outcome depends on the parameters you set and on the input (dictionary). But using this and other analyzers one gets a good idea. I especially like the one from Colemak-DH, but there is no chance to input own text (for Dutch), so that is one reason I use others as well. Anyway, it's always a trade-off and besides Polish the Minimaks layout seems pretty robust and is not the best layout for a given language but mostly in the first third, which for me is surely interesting. Other layouts like cry, mine, AdNW could also be quite good for English, German and Dutch, but move many more keys around. I'll play around for some days with the current Minimaks layout and possibly also have a look at the latter ones mentioned.