|
PureBytes Links
Trading Reference Links
|
Dennis,
I'd have to agree with Alex here. Bob Brickey may have been a genius, but
he's not going to break the laws of computational science. What you're
suggesting is akin to breaking mathematical or physical laws, like defying
gravity. I hope you can understand that I tend to be skeptical of anyone
who claims to have broken natural laws. My typical response is: show me!
(of course, if you start floating across the room it's a different story :-)
Alex mentions collection size as an important consideration. I would add
that initialization and per-unit computational costs can also be factors.
Considering all these things, we'd have to see the actual benchmark code.
Something in those numbers cries fowl. Perhaps the other initialization and
per-unit calculations suffer from poor implementation.
Of course, if the implementation works for you, great!
Kevin
At 12:46 PM 2/9/2004 -0800, you wrote:
> I think this Bob Brickey fellow made some errors, or created some code
> having huge overhead handling problems for the more efficient sorts.
You obviously don't know Bob Brickey. :-) I believe he was active here
before your time.
|