On 2/23/06, Mike Pall <mikelu-0602@mike.de> wrote:
Hi,
Chris Marrin wrote:
> http://dada.perl.it/shootout/
This page is outdated. Active development takes place here:
http://shootout.alioth.debian.org/
I actually prefer the old shootout to the new stuff being done on
alioth. The new benchmarks are so complicated that programmer
ability can have an enormous impact on the results. It was
bad enough before with the relatively simple tests but some of those
new versions are way beyond the complexity of the original
shootout. More complexity means more chances for the writer of
the code to screw up. Plus they take a lot of skill and time to
implement so many languages are missing from the results.
Don't get me wrong, I'm sure the new stuff can be
useful and everyone involved has the best intentions. It just
seems like it has taken what was a simple, neat, KISS type thing and
turned it into a bloated unreliable mess with huge gaping missing
parts. Just looking at the results now there is a strong bias
towards non-conventional languages. Except in a few specialized
cases, the old standards like C and FORTRAN should set the marks to
beat. I don't see that in the current results.
