Metaruby, BFTS, Cardinal and Rubicon - State of play?

[Austin Ziegler]

Austin, you've stepped from criticism to personal vendetta.

Not in the least. You demonstrated *then* that you were wholly
uninterested in being responsible.

You make abusive comments without any reference to what is actually
shown on the shootout website. In a similar tirade, six months ago, you
complained about a benchmark that, even then, hadn't been shown in the
shootout for at least 6 months.

Not shown?

You *do* prominently link these from the shootout home page.

Where exactly is that Perl program shown in the current benchmarks?
Here's the link http://shootout.alioth.debian.org/gp4/

http://shootout.alioth.debian.org/old/benchmark.php?test=ackermann&lang=perl&id=3

You may have changed the set of benchmarks. I'll do some research now
and see if you've actually cleaned up your methodology. I doubt it,
but if you have, I'll publicly post an apology. If you haven't, you'll
clean up your methodology. Otherwise, I'll update my criticisms, but
you won't get anything positive from me about the shootout.

It will be about two weeks before I can look at this, but I will do so.

Do we have an agreement there?

-austin

 

[Gregory Brown]

It wouldn't be necessary if you made the changes that I recommended
two years ago, so that people take what you've done a hell of a lot
less seriously, or that you take it a lot more seriously and do some
rigorous analysis and reject invalid submissions.

As it stands, you refuse to do either. So, yes, I *will* bring up the
fact that the shootout is a "silly place" every time anyone refers to
it for serious matters. It's not worth looking at because you pimp it
yet refuse to take responsibility for ensuring quality results.

Austin, I hear you. I think many people on this list hear you too. I
bet most of us agree with you. But this is incredibly off topic, and
if you *must* go on about it, please do so somewhere more suitable.

I fear your message is slowly getting lost in the amount of noise you
create each time this comes up.

 

[Isaac Gouy]

Not in the least. You demonstrated *then* that you were wholly
uninterested in being responsible.


Not shown?

You *do* prominently link these from the shootout home page.


http://shootout.alioth.debian.org/old/benchmark.php?test=ackermann&lang=perl&id=3

So you couldn't find that Perl program in the current benchmarks.

We do still show "the obsolete 2001 Doug Bagley Benchmarks", and we
*do* prominently label them "the obsolete 2001 Doug Bagley Benchmarks",
and of course they were not referred to by Kevin Tew.

You may have changed the set of benchmarks.

There's no /may/ - the benchmarks were changed a year ago.

I'll do some research now
and see if you've actually cleaned up your methodology. I doubt it,
but if you have, I'll publicly post an apology. If you haven't, you'll
clean up your methodology. Otherwise, I'll update my criticisms, but
you won't get anything positive from me about the shootout.

It will be about two weeks before I can look at this, but I will do so.

Do we have an agreement there?

Of course not.

Your opinion of the shootout "methodology" is no more than your opinion
of the shootout "methodology". When you have up-to-date information,
argue the merits of your opinion.

 

[Austin Ziegler]

So you couldn't find that Perl program in the current benchmarks.

We do still show "the obsolete 2001 Doug Bagley Benchmarks", and we
*do* prominently label them "the obsolete 2001 Doug Bagley Benchmarks",
and of course they were not referred to by Kevin Tew.

This is not the 2001 benchmark that we're talking about, Mr. Gouy.
This is the one that you had that replaced the 2001 benchmark.

There's no /may/ - the benchmarks were changed a year ago.

You snipped the context in your split of the paragraph. The
implication here is that I don't think you've changed your
methodology, which is what the problem was. Not the benchmarks
themselves.

Of course not.

Then I see no reason to research the issue aside from my own sense of
honour. I'll do it, but you have confirmed for me that I'll find
exactly what I'm expecting to find and that you haven't done anything
to improve your methodology, so you'll be getting no apology unless I
find that -- as unlikely as it appears to be -- you have actually
changed the methodology you use.

Cluebat: your methodology is and always has been the problem. Until
that improves, the shootout remains completely useless.

-austin

 

[Isaac Gouy]

Austin, I hear you. I think many people on this list hear you too. I
bet most of us agree with you. But this is incredibly off topic, and
if you *must* go on about it, please do so somewhere more suitable.

I fear your message is slowly getting lost in the amount of noise you
create each time this comes up.

Here's somewhere more suitable
http://alioth.debian.org/forum/forum.php?forum_id=999

 

[Isaac Gouy]

I've changed the subject line - my apologies for not having done so
sooner.

This is not the 2001 benchmark that we're talking about, Mr. Gouy.
This is the one that you had that replaced the 2001 benchmark.

I don't know what you mean.
Those "obsolete 2001 Doug Bagley Benchmarks" /are/ the same tasks that
appeared on Doug Bagley's old Great Computer Language Shootout. That's
simple enough to confirm, just check with the Wayback Machine archive
http://web.archive.org/web/20010124090400/http://www.bagley.org/~doug/shootout/

You snipped the context in your split of the paragraph. The
implication here is that I don't think you've changed your
methodology, which is what the problem was. Not the benchmarks
themselves.

No Austin, none of your posting was snipped, the whole posting was
shown.

Then I see no reason to research the issue aside from my own sense of
honour. I'll do it, but you have confirmed for me that I'll find
exactly what I'm expecting to find and that you haven't done anything
to improve your methodology, so you'll be getting no apology unless I
find that -- as unlikely as it appears to be -- you have actually
changed the methodology you use.

Cluebat: your methodology is and always has been the problem. Until
that improves, the shootout remains completely useless.

Are you usually given a blank check?

Anyway, I look forward to a diatribe against /the current/ Computer
Language Shootout.

As a practical matter
- it's obvious that this topic has long been tedious to some folk on
this list, so let's be considerate and give them a subject line they
can filter out

- better yet, post the diatribe in The Computer Language Shootout
Benchmarks discussion forum
http://alioth.debian.org/forum/forum.php?forum_id=999

- better yet, post the diatribe somewhere else and link to it everytime
you feel the need to make the same points
http://wiki.rubygarden.org/Ruby/page/show/BenchMarks