[Shootout-list] when benchmarks are advocacy

[Brent Fulgham]

On 2004-09-23 00:35:06 -0700 Bengt Kleberg 
<bengt.kleberg@ericsson.com> wrote:

> Brandon J. Van Every wrote:
> ...deleted
>> I think the 'main, frontlined' composite score should obey the Least
>> Common Denominator of all the languages.  It should not include any
>> tests for language-specific features, like garbage collection or
>> concurrency or thread locking or big integers or whatever.  I mean, 
>> why
>> should C++ be penalized just because Erlang has some great 
>> concurrency
>> stuff in it?  I'm not personally writing a concurrent application, I
>> don't care.

Then don't include the tests you don't like in your scoring selections 
(in the Scorecard page).
I assure you that you can manipulate the weights in such a way that 
you can make C++ win
over everyone else.  The Shootout is very flexible this way.

The shootout is designed to show the relative merits of the various 
languages.  An engineer
writing scientific software should be able to visit the site and 
determine if she should use
Fortran (not tested yet, alas) or Scheme for her work.  A lexographer 
should be able to see
which languages have the best combination of lines-of-code and speed 
with respect to
text processing.  A game programmer should be able to see what 
languages can process
blazing-fast matrix algorithms.  etc...

>> We'd also need to decide what's ok to have in a library and what 
>> must be in the
>> language proper.
> 
> how do you define a library?
> how do you define the language proper?

C and C++ are given a lot of latitude here.  I'm a bit easier than 
Doug was, in that I think
anything that is commonly used by developers is a reasonable 
candidate.  For example,
I have no problem with Python's Matrix entry using the NumPy module, 
since you would be
hard pressed to find anyone using Python that doesn't use that module.

-Brent