From: Bill Kelly <billk@...>
Date: 2002-08-10T04:30:48+09:00
Subject: Re: Coding challenge (on Ruby Garden) [long]

From: "Christoph" <chr_news@gmx.net>
[...]
> Hm, these units tests are moving targets;-). I changed the
> algorithm to cover this test case.  I also fixed editing
> mistake in the Unit tests ...

Congrats!  It passes all the tests... for now!!!  ;-)

Here's how it performed relative to a couple others from
before that still pass all the tests:

# chr      (10 iterations)    3.204 (0.3204 per iter)  nitems=60
# george   (10 iterations)   40.638 (4.0638 per iter)  nitems=60
# bwk      (10 iterations)   78.503 (7.8503 per iter)  nitems=60

Interestingly, though, although an order of magnitude faster
than those competitors, it seems to have the same O() somehow?

Because with nitems=70 george and bwk took "forever".  With yours
at 70 we have:

# chr      (10 iterations)   19.288 (1.9288 per iter)  nitems=70

But with nitems=80, yours also took "forever".  (I broke it 
after about a half hour still not finishing the 1st iteration.)

I was kind of surprised by that . . . 


Regards,

Bill