[kplato] Summary of Data Discussion

[bilbo]

On 21 Jun 2001, at 12:01, Jim Sabatke wrote:

> There is a standard way of calculating risk.  The idea is that there is 
> natural risk that is uncontrollable in any task.  Earlier emails detail 
> the standard calculations for such risk.
Ok, sorry I missed that, I'll check the archive.

>  I did make one error in the 
> presentation; risk is associated with effort, not duration.   It is 
> possible that effort and duration can be different values, that is, 
> duration can be longer because of inactive periods of effort.   Anyway, 
> providing for "frequently sick" employees is another kind of risk 
> factor.  One way to handle this would be to reduce the hours available 
> in the employee's calander.
I was talking to a friend about this last night and one of his pet 
hates with MS Project is that if he changes the duration of a task it 
changes the effort as well, he'd like to be able to hold one of these 
invariant whilst changing the other on a task by task basis.
Presumably a report to highlight the under-allocation of resource 
where the duration is reduced but the effort remains the same 
would be standard? (I don't see over-allocation being as big a 
problem but ....)

In sw projects I'm used to quoting estimates for tasks as, for 
example, 30 mandays +/- 5days. Does this translate easily into 
risk or is risk to be handled separately, and if so should we be 
including uncertainty? It could be required to enter the base 
estimate plus the positive uncertainty (35 days from above) but this 
smacks of imposing an approach and reducing flexibility.
Is it only (some) software projects that work this way - surely other 
disciplines have similar problems with estimation?

regards,
        Bill