[Shootout-list] Directions of various benchmarks

[Bengt Kleberg]

Jon Harrop wrote:
...deleted
> For data structure intensive code, no. The "nth" program from my book, for 
> example, sees comparatively tiny performance differences when compiled with 
> different optimisation flags because the programs are memory limited. The 
> performance is then a function of the algorithmic efficiency of the standard 
> library implementation. This is also something which can be productively 
> measured by the shootout.

this is a good idea.

it would mean a change of policy for the shootout. currently the 
shootout has decided to only use tests that minimise differences between 
languages. if there is sufficient interest in allowing tests to 
highlight differences we will need to rewrite existing tests.

i am very willing to update the neccessary files myself if the work load 
is too big on the poeple currently doing all the work.


bengt