doubles and cos(pi / 2)

Stephen Weeks sweeks@intertrust.com
Wed, 4 Oct 2000 13:41:07 -0700 (PDT)


> I don't understand
> how the OCAML entries got this test right, 
...
> OCAML		6.12303176911e-17

I should have mentioned that this was OCAML 2.02.  I just downloaded 3.0 and
indeed, it prints the result as zero.  So now OCAML and SML/NJ agree the result
is zero, while everyone else thinks the result is 6.123E-17.