Husk corpus

[cairdcoinheringaahing]

Similarly to this and this issues opened in the Jelly repo, I decided to run Lynn's method on Husk answers (with a few modifications).

SEDE query used to get all Husk answers on PPCG. Code used to analyse the data. Final corpus:

2-graphs    3-graphs    4-graphs
  59 fo      16 o=¹       9 fo=¹
  50 mo      15 fo=       9 o=¹Σ
  39 o=      15 S=↔       7 …"AZ
  36 S+      14 R'        7 ḟo=⁰
  36 '       12 `:'       7 o=⁰Σ
  35 S=      11 o=⁰       6 S+s"
  33 mΣ      11 100       6 o`:'
  32 İp      11 ḟo=       5 oS=↔
  31 U¡      10 =¹Σ       5 ←LU¡
  30 mȯ       9 LU¡       5 z*İ_
  30 !¡       8 T'        5 mo`:
  30 fȯ       8 =⁰Σ       5 ↔T' 
  30 R'       8 S+s       4 mLgp
  27 10       7 ▲mL       4 mΣ∂Ṫ
  24 :'       7 oS=       4 Σ∂Ṫ*
  22 Ẋ-       7 mLg       4 Sze↔
  22 ▲m       7 `-N       4 §&o=
  22 mL       7 fo¬       4 "S+s
  22 C2       7 *10       4 fS=↔
  22 İf       7 …"A       4 foS=
  21 =⁰       7 "AZ       4 o=¹Π
  21 ¡o       7 fS=       4 "()"
  21 00       7 (mo       4 ḟo=¹
  20 ṁo       7 S=u       4 =¹ΣṖ
  20 `:       6 Lgp       4 mR' 
  19 !f       6 Σ↑_       3 fo=⁰
  19 ḟo       6 S+o       3 U¡ṙ1
  19 Σz       6 ze↔       3 	
  19 =¹       6 !İp       3 !uΣz
  18 ¡ȯ       6 ↑İp       3 ΣTC2

I'm not familiar enough with Husk to offer any informed suggestions or remarks, but I hope this helps!

[barbuz]

I would rather go with some form of scientific notation rather than a single builtin for 100. Seeing how common the "00" 2-graph is, 100 is not the only problem.

[cairdcoinheringaahing]

The corpus analyser does work on overlapping matches, and doesn't care about matching part of a string in the "smaller" groups, so from the numbers we can see that 100 appears 11 times, and 00 21 (note that this isn't in 21 Husk programs, but that instead, of all Husk programs on CGCC there are 21 (potentially overlapping) occurrences of 00.

Nonetheless, a builtin for 100 that takes less than 3 bytes would likely be a good idea, as would a single byte for a space character, and some way of shortening fo=¹, either through assigning fo= to a single byte or some other way

(I've learnt more Husk since opening this issue :))