BUILDING

~ cabal install

DICTIONARY

- currently just hardcoded to /usr/share/dict/scrabble
- recommended to grab owl scrabble list:
  ~ wget http://dzl.no-ip.org/g/scrab/owl.txt.bz2
  ~ bunzip2 owl.txt.bz2
  ~ sudo mv owl.txt /usr/share/dict/owl
  ~ sudo ln -s /usr/share/dict/owl /usr/share/dict/scrabble
- or use another list that you prefer

GAMEPLAY

- type out words
- space shuffles letters on rack
- enter submits a word
- escape goes to the next round
- ctrl+c quits

DEVELOPMENT

todo
- switch to rand monad
- switch to bytestring-trie, or see if that's faster?

OLD NOTES

- name ideas
  - charism
  - chorizo
  - karst
  - karstic
  - lexis
  - lexemic
- binary encoding Trie was not worth it
- bytestring load cuts time from about 3s to 2s but limits us to ascii

- penalty for wrong words?  (count bad ones, or have timer with time deduction)

- restrict to high point words?  currently studying lots of common-letter small
  words (less useful..?)

- wn_i := number words wide of i-letter words section
- w := total width
- h := total height
- h_i := height of section of i-letter words
- w_i := width of section of i-letter words
- n_i := number of i-letter words

w = max_i w_i
h = n - 1 + sum_i h_i
h_i = ceil(n_i / wn_i)
w_i = wn_i * i + wn_i - 1
=> wn_i = (w_i + 1) / (i + 1)

so
h ~= n - 1 + sum_i n_i (i + 1) / (w_i + 1)
  ~= n - 1 + sum_i n_i (i + 1) / (w + 1)
   = n - 1 + (n + sum_i i*n_i) / (w + 1)

w = (n + sum_i i*n_i) / (h + 1 - n) - 1

e.g. n = 4 and w = 10
h = 2

w = n (n + 1) / h - 1


aa ab ac ad
ae af ag ah
ai
aaa aab aac
aad aae

w = 11
w_1 = 11
w_2 = 11
h_1 = 3
h_2 = 2
n_1 = 9
n_2 = 5
wn_1 = 4
wn_2 = 3
n = 14
h = 5


h_i = ceil(n_i / wn_i)
3 = ceil(9 / 4) YES
2 = ceil(5 / 3) YES


wn_i = (w_i + 1) / (n_i + 1)
i = 1: 4 = 12 / 10 no

h = sum_i ceil(n_i / wn_i)
  = 3 + 2

 ~= sum_i n_i / wn_i
  = 9 / 4 + 5 / 3 = 47 / 12 = just under 4

 ~= sum_i n_i (n_i + 1) / (w + 1)
  = (n + sum_i n_i^2) / (w + 1)
  = (14 + 106) / 12 = 10  NO


h = (n + 1) sum_i n_i / (w_i + 1)
5 = 15 * (9 / 12 + 5 / 12)   NO