Binary size - compress word lists? about toipe HOT 3 OPEN

How about storing it as a trie or Radix Tree

from toipe.

@Samyak2 I think my suggestion is a little wrong, even though a trie or radix tree might be faster in computation of the random word(not entirely sure of this), If im not wrong, what you want to do is find a way to store the word lists by reducing the size of the file

Taking inspiration from the trie maybe you could store words in the following fashion
where the number represents the number of letters taken from the 0th word and the letters represent the letters to be added
A example taken from below would be <5,t> which amounts to applet by taking the first 5 letters from apple
Note:- // is used a comment

0, apple
5, t  // applet
3, ear //appear 
0, ball
3, d //bald
...
...

Now one drawback in storing it this way would be the extra computation of finding the most recent '0'th word in the series, though it might be that heavy it would be nice to avoid it altogether. A variation of the above in storing it could be

0, 0, apple
5, 1, t
3, 2, ear
0, 0, ball
3, 1, d

Where the second integer represents number of steps to the 0th word

And as for random words in tries this might help

from toipe.

How about storing it as a trie or Radix Tree

@lucasace that's a good way to compress the data. Although, I'm not sure if it will help in selecting random words in a uniform manner from the list. Do you have any ideas about implementing that in a trie or a radix tree?

from toipe.