[Shootout-list] Threads and Threads-Flow Proposal

[Isaac Gouy]

> however, if you, or somebody else for that matter, are very
> persistent on this list, you might eventually be told what the narrow
> focus of the shootout will allow in the way of improvments.

No, that happens quite quickly, like this - read the FAQ.

As in any other endeavour, if you really want to make a difference then
do all the work! Make an excellent case why change is needed *and*
provide a complete implementation.

So the way for Jabari Zakiya to make his suggested minor change to
heapsort a reality (a minor change which at face value will have no
effect on the relative performance of the programs), is to rewrite
*all* the heapsort programs himself - instead of pushing a pile of
make-work at volunteers.

The way for Kevin to make his suggested improvement to threads-flow a
reality is to measure half a dozen languages across 3000 300 30
threads, make the case that the change is neutral, and provide
implementations of all the threads-flow programs working at 300. Or he
can file a Feature Request and wait for someone else to get to it.

The way for James to make his strict notion of only one program per
language per benchmark a reality is to make an excellent case for why
it matters - sensibly, he already offered to do the work.

Bengt, it doesn't seem that you're trying to achieve anything in
particular, so all I can really suggest is that there are 10 benchmarks
without Erlang implementations.

best wishes, Isaac


__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com