NoFib for benchmarking vs testing

[Robert Ennals]

[snip]

All sounds sensible.


What I'm really suggesting is that, unlike the other programs, I don't al=
ter =

these programs to make them take approx-ten-seconds to run, as there does=
n't =

seem much point.

We thus have programs grouped into "those that take 10 seconds, and are u=
seful =

runtime benchmarks", and "other programs".



If there is useful information to be gained from them, then great, but it=
 =

probably isn't particularly useful to try and distort them into running f=
or 10 =

seconds.



-Rob