From: gotoken@... (GOTO Kentaro)
Date: 2000-12-30T04:30:16+09:00
Subject: [ruby-talk:8292] Re: speedup of anagram finder

In message "[ruby-talk:8287] Re: speedup of anagram finder"
    on 00/12/30, "SHULTZ,BARRY (HP-Israel,ex1)" <barry_shultz@hp.com> writes:
>The idea is this: use the math theorem that upto order of the prime factors,
>'every positive integer can be written uniquely as a product of primes'. To
>create an integer key for each word, I first setup a hash which maps ascii
>codes for a-z and A-Z to the first 26 primes. Then as I each_byte the word,
>the key is built of the product of the
>primes associated with each letter in the word.

Neat!  It is a kind of Goedel number. I've never seen any practical
application of this numbering scheme. 

>bm do |x|
>  GC.start
>  x.report("primes"){barry(wordlist)}
>  GC.start
>  x.report("joe"){joe(wordlist)  }
>  GC.start
>  x.report("devel"){devel(wordlist)  }
>end
>
>      user     system      total        real
>primes  1.923000   0.000000   1.923000 (  1.923000)
>joe  2.533000   0.010000   2.543000 (  2.543000)
>devel  2.404000   0.010000   2.414000 (  2.413000)

You can justify output lines by "bm(10) do |x| ...". 

-- Gotoken