I think I would ignore those non-working hours and just consdier that a task
the takes 1 day represents 8 hours of work, even though those hours are
really not being done at 100% efficiency. My reasoning is based on where
you most likely come up with duration estimates in the first place. Let's
say the task in this project is to polish 1000 fids - how do I know how long
that should take? Most likely I'll look back on what we've done in years
past and find that in last year's project it took us 10 hours to polish 100
fids. So for this year I'll probably assume it will take us about 100 hours
to polish the 1000 required this go araound, 10 time more. But when we
*did* that work last year the people doing it spent some time smokin' and
jokin', checking email, going potty, etc. So that 10 hour duration we have
in our records actually is the total of the productive time and the time
lost to non-productive activities. If we use it to estimate this year's
task duration we can probably assume the same proportion of lost time as we
had before and so that 100 hour we come up with already takes into account
we're really getting 75% efficiency. Since it's not our job as PM's to
micromanage our resource's day, if we just ignore the fact that "8 hours of
work" in our project plan is really 6 hours procuctive time and 2 hours
non-productive time, our schedule will work out just fine.