[lnkForumImage]
TotalShareware - Download Free Software

Confronta i prezzi di migliaia di prodotti.
Asp Forum
 Home | Login | Register | Search 


 

Forums >

comp.lang.ruby

Re: python/ruby benchmark(don't shoot the messenger

Nuralanur

6/13/2005 9:54:00 PM

Tanner Burson wrote:
> I've tried to stay out of this thread, because I don't feel it's
> providing, or furthering much intelligent discussion. But what on
> earth are you attempting to say by this quote? Or are you just
> throwing out random things for the sake of saying that somewhere, for
> a given problem domain, there is a language more suited than Ruby? If
> so thanks, point taken. If not would you mind clarifying by saying
> something a bit more descriptive? Is the above quote actually in
> reference to a given language you're attempting to introduce everyone
> to? Or is it just a generic example with no meaning whatsoever?

Sorry.

Only 2 options were presented - slow-to-write fast-to-run C and
fast-to-write slow-to-run Ruby - there are other possibilities. That's
all.

Yes, of course. I was stating my personal views without claiming
generality.
Mmh, benchmarks seem to be good essentially for one thing, to judge
by this thread - propagate bad mood. I tried to change that a bit , without
success apparently.
If people should really be running away from Ruby or see
it as a language good for web-development only because of bad
benchmarks, my proposal would be to try to identify a small set
of features which are slowing down the relative performance of Ruby in
applications many people feel important and enhance them in a future version.
Maybe that's what this not-so providing thread could eventually do.
So far, I myself am quite happy with Ruby's performance, so I couldn't
open a list, but nevertheless, I think that if it is was possible to identify
some consensus about say the five biggest downslowers in important
applications
and fix them in a later version, this could achieve more than trying to
put much more work into fixing many more problems.

Best regards,

Axel
2 Answers

Ryan Leavengood

6/13/2005 11:19:00 PM

0

My problem regarding these (or any) benchmarks is that a language could be
tweaked and designed to satisfy the benchmarks, whereas real world
performance may not be that good. It was mentioned how NVidia did this
with video cards. I imagine this is part of the perspective that Austin is
coming from.

These are very specific, pedantic benchmarks that don't seem to allow any
real creativity in implementation. The power or elegance of many of the
languages is lost in implementing an identical algorithm to solve a
certain problem. Surely varied algorithms that give the same answer should
be allowed to show the power and semantics of the given languages?

I'm not as passionate about this as Austin, but I certainly see where he
is coming from.

Ryan


Isaac Gouy

6/13/2005 11:37:00 PM

0

Ryan Leavengood wrote:
> Surely varied algorithms that give the same answer should
> be allowed to show the power and semantics of the given languages?

Surely that would show the power of algorithms. (Yes, memoising
Ackermann is much much faster.)