[Shootout-list] Re: Integer overflow, massivly parallell

Brandon J. Van Every vanevery@indiegamedesign.com
Fri, 8 Oct 2004 11:29:40 -0700


Isaac Gouy wrote:
> Bengt Kleberg wrote:
> > Isaac Gouy wrote:
> > ...deleted
> > > A good test will rank all the languages. There will be people who
> > > are more interested in finding out if X is better than Y than in
> > > proving Erlang is best.
> >
> > from your silence i gather that you have no furhter
> > interest in this
> > discussion. i find this unfortunate, but respect your decision.
>
> imo The discussion was stuck on disagreement about what makes a good
> test:  is it enough to show the "best" language or should the
> test rank all the languages?

I agree firmly with the general principle that a good test ranks all
languages.  Performance is but one factor.  People are always implicitly
evaluating many other factors when they look at a performance spread,
i.e. availability of implementations, libraries, suitability to problem
domain, what they know already, etc.  Thus I want to know who is 1st,
2nd, 3rd, 4th... at performance, not just who is 1st.  I want to know
how far apart the leaders and the followers are.  In engineering it is
often the 80% solution that wins.


Cheers,                         www.indiegamedesign.com
Brandon Van Every               Seattle, WA

20% of the world is real.
80% is gobbledygook we make up inside our own heads.