[Shootout-list] Stuff

[Brent Fulgham]

--- Bengt Kleberg <bengt.kleberg@ericsson.com> wrote:
> Isaac Gouy wrote:
> ...deleted
> > 
> > While we're speculating, does anyone want
> > implementations using different algorithms with
> > different running times?
> 
> yes, i do. _if the benchmark design allows it_.

The benchmark allows this using the 'interesting
alternative' functionality.  The only limitation
is that one of the 'n' solutions must be marked as
the 'real' solution for purposes of the CRAPS
table.

> note: it should be possible to improve the design if
> flaws are found. the worst thing that can happen is 
> to have a flawed design, and a word-of-mouth 
> convention that does not allow certain solutions.

I am trying to address problems as we discover them.
Unfortunately, I can only provide a few cycles a day
to these tasks, so they often take weeks to get done.

It would be much better if the shootout machine
lived on the network such that others could log onto
it and update the scripts, etc., but I do not have
such a connection.

A while back I posted a plea for someone (University
perhaps) to host the shootout on a network-accessible
box, but this has not been provided.

Until then, the shootout 'runs' remain very much a
one-man show. :-(

I've tried to address this to some extent by having
the build machine log onto the net, download CVS,
and build every hour or so, but I can't keep the
machine running round-the-clock (due to noise
problems), so it's generally just a few hours a day
that this happens.

-Brent