FAQ

Compare Canary to Dusk is not fair!

Yes, I agree, but I used to main Canary-ortho, it makes sense for me to compare them.

For comparison with other thumb layouts, check out Comparison.

Why not swap 2 index block?

It increases alternation by a whopping 4% to 32.XX% on Cmini with Monkeyracer corpus, which is too high for my taste.
OW and WO middle-index full stretch would suck. Swapping W and Q won't help because then WA would suck.
The intended CY, YC, PY, YP alts would be less feasible. Think words like happy or privacy.

How much worse is vowel thumb space and can I use thumb R with vowel thumb?

All of the below still stands true, but I have some new perspective on this here.

It is commonly known that space should be pressed with the consonant thumb because it has lower redirection and higher roll, and so why is the thumb R in Dusk on the consonant side?

The reason is simple - thanks to a200 analyzer, I've come to the conclusion that even after considering the extra 0.52% redirection incurred by vowel space thumb, having R on the consonant thumb as opposed to the vowel thumb leads to approximately 3.87% lower redirection overall.

But then you may wonder: Is the redirection caused by thumb even real? How come that I have not felt any space redirection?

For me at least, after having tried Dusk with vowel thumb R for a few days, most redirection involving R is just as good, if not better than, the redirection involving index finger, but with one exception: when index L is also involved. Take the word earlier for example, that word just feels bad to type with vowel thumb R.

What is the True SFB if I alt and slide?

If you

Alt PY, YP, CY, 'M, and 'L.
Slide UE, OA, HL, YW, and FS.

then SFB is almost none, leaving really only GY, LM, I., and I,. None of which is common (unless you type almost a lot).

Can I re-arrange index letters?

Yes - arrange them how you like them!

Can I use `Dusk` on a rowstag keyboard?

It is possible to do so with wide mod, but Dusk was designed for colstag keyboards, so I can't say the experience would be good. You may or may not want to re-arrange some letters.

Is Dusk good for e200 on MonkeyType?

It is :D

It doesn't have any SFB (0.00%) on e200 and with remarkably low SFS as well.

dusk
  x f d p q  j ' o u .
  n s t c y  m h a e i
  b v k g w  z l _ / ,
          r                 

ENGLISH-200:
  Alt: 30.73%
  Rol: 50.11%   (In/Out: 21.60% | 28.51%)
  One:  1.34%   (In/Out:  0.67% |  0.67%)
  Rtl: 51.45%   (In/Out: 22.27% | 29.18%)
  Red:  3.79%   (Bad:     0.22%)

  SFB: 0.00%
  SFS: 3.34%    (Red/Alt: 1.56% | 1.78%)

  LH/RH: 47.04% | 52.96%

`W` position... Really?

There is actually another version of Dusk: Dusk_optimized.

To get rid of W inner key (and optimize SFS and SFB), Dusk_optimized moved P to be with BN, and in doing so, it gets rid of the inner corner key and achieves even lower SFB and SFS. However, the cost of it is a whopping 1.41% increase in half scissors according to Keysolve.

I hesitate to recommend Dusk_optimized for the following reasons:

Words like prob, problem, public being 2 row ring jump (much more freq than f_v or f__v on Dusk).
Frequent nk half scissor.
Worse ng and g_t.

dusk_optimized
  z p d g q  j ' o u .
  s n t c y  m h a e i
  f b k w v  x l _ / ,
        r                   

MONKEYRACER:
  Alt: 29.10%
  Rol: 49.88%   (In/Out: 21.12% | 28.76%)
  One:  1.69%   (In/Out:  0.65% |  1.05%)
  Rtl: 51.57%   (In/Out: 21.77% | 29.81%)
  Red:  3.48%   (Bad:     0.39%)

  SFB: 0.46%
  SFS: 3.81%    (Red/Alt: 1.12% | 2.69%)

  LH/RH: 46.91% | 53.09%

How did you make the figures in Layout?

It is from here with the following settings:

layout:
  qmk_keyboard: chocofi
  #qmk_layout: LAYOUT_split_3x5_2
layers:
  DEF:
    - [X, F, D, P, Q, J, "'", O, U, .]
    - [N, S, T, C, Y, M, H, A, E, I]
    - [B, V, K, G, W, Z, L, "_", /, ","]
    - {}
    - {t: R}
    - {}
    - {}
    - {t: Space}
    - {}

Dusk Keyboard Layout

FAQ