Half a croissant, on a plate, with a sign in front of it saying '50c'
h a l f b a k e r y
Normal isn't your first language, is it?

idea: add, search, annotate, link, view, overview, recent, by name, random

meta: news, help, about, links, report a problem

account: browse anonymously, or get an account and write.

user:
pass:
register,


               

finite differences

Is it already used as an indexing method?
  (-1)
(-1)
  [vote for,
against]

Take a word, say "word", list the ascii codes of its characters, say {119, 111, 114, 100}. Now take the first difference {-8, 3, -14}, followed by {11, -17} and finally {-28}. So "word" is sponsored by the numbers 4 (its length) and -28 (its 4th finite difference). Now set up a search engine that seeks out and collects 4 letter words, another that collects words with nth finite difference of -28, and force them to share their intelligence with each other.
LoriZ, Aug 07 2001

[link]






       bena berm bevy cake cern chun cise club dali fest fire fish flue gall hers idle keld love mako moyo pahi pail pipi pouf rain rine rist rote royt sell slur Tape temp ting tora vair velo viol waeg waky whop wild word   

       I fail to gain insight from this, but maybe you can sell it to a numerologist.
jutta, Aug 07 2001
  

       First Law of Feng Shui: stays in its element.
reensure, Aug 14 2001
  

       Forced intelligence is just wrong.
The Military, Aug 14 2001
  

       jutta: I think you misplaced your annotation... it obviously belongs in The Greatest Story Ever Told. ;)
PotatoStew, Aug 14 2001
  

       I don't think so. The number of possible final differences is very low and there are a lot of four letter words. That technique seems anyway about the same as assigning codes to a limited number of dictionary entries and just using those. If you wanted to avoid losing information, you'd have to represent all the differences as well as the initial value -- which would, I think, actually take an extra bit per character for any range of codes. You might be able to identify fancy patterns using differences and represent those -- but then you'll be trying to do something that any ordinary archiving utility already does pretty well.   

       (I don't know much about these things either, though, unfortunately.)
Monkfish, Aug 15 2001
  

       I had assumed that jutta's list all compressed down to the same value, but I didn't bother checking to see if that was actually the case.
PotatoStew, Aug 15 2001
  

       Looking at it again, it turns out that the original idea actually depends on there being lots of words with the same length and final difference to... to carry out its purpose.   

       (The other problem for even long words, of course, is that you'd have to go through lots of possibilities for a given length and final difference before you found a known word (which would require you to have a list of all words handy, anyway). I guess you could then store the word-difference/length information for next time, but it still doesn't seem very sensible in a world bursting with good ways of compressing text. Not that that's in any way the point.   

       Won't the "minimum set" idea always be inferior to just dropping letters off the end and trying to match the word from a list? You'll have the same information loss and it will take more space (twice the range for each character) and more effort to find the word. You'll also probably destroy the fine numerological relationships of the letters. On a more positive note, there's a good chance I'm wrong.)
Monkfish, Aug 15 2001
  
      
[annotate]
  


 

back: main index

business  computer  culture  fashion  food  halfbakery  home  other  product  public  science  sport  vehicle