News:

Pretty crazy that we're closer to 2030, than we are 2005. Where did the time go!

Main Menu

Interesting Problem

Started by iago, December 06, 2005, 11:48:31 AM

Previous topic - Next topic

0 Members and 2 Guests are viewing this topic.

Sidoh

Quote from: Nate on December 08, 2005, 07:21:51 PM
Maybe you guys should start with something easier like sudoku's then try this.

Also you want to treat this kinda like finding collisions, find all possbilities for each word, then compare the letter = letter values between words to find what works then out but all possible working combos

Seems kind of inefficient to me.  Oh well.

An equally interesting problem (in my opinion) would be a Rubix Cube solver.  I was messing around with one of those on the internet.  It would probably be a bit more complicated since you have to translate numbers into motions.

Nate

Actually its probably the most logical.  Take the first word and use a dictionary to find all possible words it could be create multiple schemas, take the second word find all possiblites and find the schemas those words fit in, which will start eliminating schemas and eventually you will narrow it down to only a few possible matches or maybe just one at which point you run it through some kind of grammar check to make sure its intelligable solution.

iago

Yeah, that sounds like a good idea. 

But a twist: don't pick the first word, pick the longest one.  You can probably narrow down the longest one more quickly. 

Nate

I like that, it would produce less possibilities faster.  I would think the word with the most letter repetition might work even better though.  It would be interesting to do it many different ways and see what would work best; longest to shortest, shortest to longest, most individual letters to most repetition, most repetition to most individual letters, and so on.

Sidoh

Quote from: Nate on December 08, 2005, 09:19:43 PM
I like that, it would produce less possibilities faster.  I would think the word with the most letter repetition might work even better though.  It would be interesting to do it many different ways and see what would work best; longest to shortest, shortest to longest, most individual letters to most repetition, most repetition to most individual letters, and so on.



Just kidding.

But seriously, yeah.  Just because it's the most logical doesn't mean it's the most efficient solution.  Then again, it could.  Who knows?!

I like the Rubix Cube solver better! :P

iago

Quote from: Nate on December 08, 2005, 09:19:43 PM
I like that, it would produce less possibilities faster.  I would think the word with the most letter repetition might work even better though.  It would be interesting to do it many different ways and see what would work best; longest to shortest, shortest to longest, most individual letters to most repetition, most repetition to most individual letters, and so on.

I think you're right.  We probably need a balance between them.  Longer + more repeated letters, figure out the best way to weight it.  Then find a way to throw out proper names. 

This is seeming possible, but still tricky.  Who's up for the challenge?

Warrior

DUDE Sudokus FUCKING own. My friend can do a whole one in 5 minutes or less. It's like amazing.
One must ask oneself: "do I will trolling to become a universal law?" And then when one realizes "yes, I do will it to be such," one feels completely justified.
-- from Groundwork for the Metaphysics of Trolling

rabbit

I can do the 3x3's in 5 minutes, it's not hard.  The 4x4 and 5x5 are fucking nuts though....

Nate

I can do a 9x9 in a couple of minutes.

rabbit

Wow, odds are you can't at all.  Sudokus are refered to by a single block (IE: square of 9 (in a 3x3)).  A 3x3 has 81 blocks.  A 9x9 would have 1369 blocks, which I doubt you can do in a couple minutes, much less a week.

Nate

I didn't know how you nummbered the blocks.  Fine i can do a 3x3 in a couple minutes.  A 9x9 would be do-able in a couple of hours.

iago

#26
Ok, here's today's (or some recent day's.. it was sitting on the table) cryptoquote:

ORIU FUZFNFUP QGLJVUK.  WG
UGN NILJN NRI JVXVNI OFNR
V PEIVN ZVEFINK GS
BURIVXNRSBX WVFUNFID
-- QVNRIEFUI HIIQRIE

I think the best would to start with would be the second one, FUZFNFUP.  How many words could have the same letter in those places, plus the 2 U's? 

I agree, it seems like the #1 thing to do is to find the word with the most repeated letters, like that. 
Second step is to take the word that that helps us with the most, and so on.  WVFUNFID would be half solved after solving the second word, and so on. 

I can provide a dictionary if somebody else wants to work on a program that can solve this :)

<edit> Another good starting word might be BURIVXNRSBX..