On Sat, Sep 11, 2010 at 7:10 AM, michael rice <nowgate@yahoo.com> wrote:

Hi Baz,

That's quite an analysis, one I'll keep for future reference.

So, my original coding was the fastest. Guess I should stop second guessing myself. ;-)

I think Bas's point was actually that you should second guess yourself and use Criterion/ghc-core if you want to know which way is fastest. On the other hand, these are microbrenchmarks and should not be taken as representative of how they will perform inside a larger application.

Evidence is always nice to have :)

Jason