In article <42B88EEC.1040005 / comcast.net>,
Glenn Parker  <glenn.parker / comcast.net> wrote:
>Ryan Leavengood wrote:
>> 
>> In my informal tests, Florian's version is already about 70 times faster
>> than the original one. It is also about 3.5 times faster than the naive
>> Java version.
>
>Big frickin' deal!  Come on, guys, you are all smarter than that.
>
>Recode the Java version using the same algorithm and *then* compare. 
>It's not like Java won't benefit just as much from the algorithmic 
>improvements.
>

Indeed it will.  However, I think the point was that with just a little 
bit of thought the Ruby version can outperform a 'first cut' Java 
implementation.  If the OP considered the original Java version 'fast 
enough' and considered the original Ruby version to be unacceptably slow, 
well a little rethinking of the algorithm and the Ruby version ends up 
being faster than the Java version which was 'fast enough'.  If he really 
needs the Java version to be faster, well, now he can go back and recode it.


Phil